Our objective was to test the hypothesis that the association between elevated luteinizing hormone (LH) concentrations and miscarriage is mediated via an effect of LH on the maternal environment, rather than on the oocyte. The impact of maternal age, ovarian function, previous IVF attempts, therapeutic (buserelin) and hormonal (LH, oestradiol, progesterone) effects occurring on the day of zygote intra-Fallopian transfer (ZIFT) or embryo transfer, and of oocyte or embryo numbers, whether they were fresh or frozen, and their mode of transfer on the occurrence of pregnancy and miscarriage following ovum donation (n = 57) were investigated. The cycles were divided by outcome into non-pregnant (n = 26), miscarriage (n = 19) and normal term pregnancy (n = 12). The circulating concentrations of LH were greater in miscarriage cycles (P = 0.046) and cycles ending in pregnancy (P = 0.04) than in non-pregnant cycles, while the concentrations of progesterone were greater in non-pregnant (P = 0.029) and miscarriage (P = 0.015) cycles than in cycles ending in pregnancy. Frozen embryos were used more frequently in non-pregnant compared to cycles ending in pregnancy (P = 0.016). Multiple regression analysis was used to investigate which factors are associated with miscarriage and identified progesterone concentrations at the time of transfer as being the only significant variable (r = 0.48, F = 8.5, P = 0.007).(ABSTRACT TRUNCATED AT 250 WORDS)