The aim of the study was to assess patient, treatment and cycle characteristics in relation to the risk of multiple conception following ovulation induction in order to reduce the prevalence of this complication of treatment. We performed a retrospective analysis of 208 pregnancy cycles achieved in the Middlesex Hospital outpatient fertility unit. These pregnancies were achieved in 175 anovulatory women who conceived after gonadotrophin or pulsatile GnRH therapy. The multiple conception rate was 13.4%. After spontaneous reductions and abortions the multiple delivery rate was 9.6%. Clinical features associated with an increased risk of multiple pregnancies were the presence of polycystic ovary syndrome and secondary infertility. Comparison between different protocols of ovulation induction revealed no relationship with the risk of multiple conceptions. Although total number of follicles was increased in the multiple conception cycles, the distribution of follicles according to their diameter on the day of human chorionic gonadotrophin (HCG) administration was similar in multiple and singleton conception cycles. Thus, the risk of multiple conception could not be attributed to an increased number of follicles of any particular size but directly related to the total number of the cohort follicles ( > or = 14 mm) and leading follicles > or = 17 mm), rising from 7% with one follicle to 33% with six or more follicles. As we could not find a specific pattern of follicular development that could be associated with multiple conception, we conclude that the difference in the ovarian response leading to multiple conception is quantitative rather than qualitative. The data presented enable the assessment of the risk of multiple conception in any given cycle.