BACKGROUND Until recently, the main source of exogenous follicle stimulating hormone (FSH) for therapeutic use in infertility had been the urine of postmenopausal women. New developments have resulted in the production of FSH in vitro by recombinant DNA technology. The extremely high purity and batch-to-batch consistency of recombinant FSH (rFSH) make it an attractive alternative to urinary FSH (uFSH). OBJECTIVE To conduct a systematic review and meta-analysis of randomized trials comparing the effectiveness of rFSH with uFSH in ovarian stimulation protocols in in vitro fertilization (IVF) or intra-cytoplasmic sperm injection (ICSI) treatment cycles. METHODS Search strategies included on-line searching of the MEDLINE and EMBASE databases from 1985 to 1999, hand searching of bibliographies of relevant publications and reviews and abstracts of scientific meetings, peer consultation and contacting the pharmaceutical companies that manufacture the gonadotropins under consideration. METHODS Randomized trials comparing rFSH with uFSH for ovarian stimulation in IVF or ICSI treatment for infertility. METHODS The main outcome measure was clinical pregnancy per cycle started. Also considered were clinical pregnancy per cycle reaching oocyte retrieval and per cycle reaching embryo transfer (ET), ongoing pregnancy per cycle started, spontaneous abortion, multiple pregnancy, ovarian hyperstimulation syndrome (OHSS), number of follicles and serum estradiol level on the day of human chorionic gonadotropin administration day, total dose of FSH, and number of oocytes retrieved. Common odds ratios (OR) and risk differences for rFSH relative to uFSH were calculated after testing for homogeneity of treatment effect across all trials. The fixed effects model was used, unless significant heterogeneity was present, in which case the random effects model was used. RESULTS The overall odds ratio for clinical pregnancy per cycle started was 1.21 (95% confidence limits (CL) 1.04,1.42) for rFSH compared to uFSH. The risk difference was a 3.7% (0.8,6.7) absolute increase in clinical pregnancy rate with rFSH. The OR for ongoing pregnancy per cycle started was 1.29 (1.08,1.54). There was no significant difference between rFSH and uFSH in the rates of spontaneous abortion, multiple pregnancy or OHSS. The total dose of FSH was lower by 406 (185,627) IU with rFSH, but there was no significant difference in the number of follicles or serum estradiol on hCG day or in the number of oocytes retrieved. CONCLUSIONS This review has demonstrated a statistically significant increase in clinical pregnancy rate with rFSH compared to uFSH, when used for ovarian stimulation in assisted reproduction. This benefit was observed only in standard IVF cycles and not in cycles in which ICSI was used.