BACKGROUND: Semen preparation techniques for assisted reproduction, including intrauterine insemination (IUI), were developed to separate the motile morphological normal spermatozoa. Leucocytes, bacteria and dead spermatozoa produce oxygen radicals that negatively influence the ability to fertilize the egg. The yield of as many motile, morphologically normal spermatozoa might influence treatment choices and therefore outcomes. OBJECTIVES: To compare the effectiveness of gradient, swim-up, or wash and centrifugation in subfertile couples undergoing intrauterine insemination (IUI) on clinical outcome as well as on semen parameters. SEARCH STRATEGY: We searched the Menstrual Disorders and Subfertility Group's trials register (30 June 2003), MEDLINE (1966 to July 2003), EMBASE (1980 to July 2003), Science Direct Database (1966 to July 2003), Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library Issue 3, 2003), National Research Register (2000 to Issue 2, 2003), Biological Abstracts (2000 to June 2003), CINAHL (1982 to July 2003) and reference lists of relevant articles. We also contacted experts and authors in the field. SELECTION CRITERIA: Parallel randomised controlled trials (RCTs), comparing the efficacy of semen preparation techniques used for subfertile couples undergoing IUI, were included. RCTs or split sample studies examining semen parameters after different semen preparation techniques were also included. DATA COLLECTION AND ANALYSIS: Two reviewers independently assessed trial quality and extracted data. Study authors were contacted for additional information. MAIN RESULTS: Two randomised controlled trials comparing clinical outcomes, including 81 participants in total, were included in the meta-analysis (Dodson 1998-I/ II; Xu 2000-I/ II). Both studies compared swim-up technique versus gradient technique in 65 subfertile couples undergoing IUI. One study compared the effectiveness of both techniques with wash technique. No trials reported the primary outcome of live birth. There was no statistically significant difference between pregnancy rates (PR) for swim-up versus gradient / wash centrifuge (Peto OR 0.55, 95% CI 0.17 to 1.76; Peto OR 1.74, 95% CI 0.2 to 14.9; PR/ couple swim-up 20%, gradient 40%, wash 12.5%) or gradient versus wash centrifuge (Peto OR 4.01, 95% CI 0.82 to 19.56; PR/ couple swim-up 15%, gradient 20%). There was no significant difference in the miscarriage rate (MR) per couple between either of the three treatment groups in the one trial reporting this outcome (MR/ couple swim-up 0%, gradient 10.3%, wash 0%. MR/ pregnancy gradient 30.3%). There was no statistically significant difference in the multiple pregnancy rate (MPR) per couple between either of the three treatment groups in Dodson 1998-I/ II (MPR/ couple swim -up 0%, gradient 0%, wash 6.3%). One triplet pregnancy was recorded. Fifteen studies comparing semen parameters after processing were included. Two studies were included in the meta-analysis, we were not able to pool results. REVIEWERS' CONCLUSIONS: There is insufficient evidence to recommend any specific preparation technique. Large high quality randomised controlled trials, comparing the effectiveness of a gradient and/ or a swim-up and/ or wash and centrifugation technique on clinical outcome are lacking. Further randomised trials are warranted. Results from studies comparing semen parameters may suggest a preference for gradient technique, but firm conclusions cannot be drawn and the limitations should be taken into consideration