[Intrauterine insemination: state-of-the-art in humans].

Despite its being used for a long time, intrauterine insemination (i.u.i.) remains debated as to its precise place and efficacy among assisted reproductive technologies. Data issued from the French Health Ministry inquiries are strictly limited to the number of cycles and the pregnancies and births including the multiple ones. Concerning 2000, more than 44,000 cycles were registered with 8% deliveries per cycle and 12% multiple pregnancies. Apart from the cervical female infertility which is considered to have the best prognosis with i.u.i., literature data remain controversial with male and unexplained infertility. Prospective randomized studies are rather scarce, particularly when considering the inclusion of untreated control population. Meta-analyses have been published for ten years, which allowed to better define the place of i.u.i. in patient management. However one may notice that the sperm cut-off parameters for male infertility and the respective contribution of i.u.i. and ovulation treatment do not allow develop some evidence-based guidelines for i.u.i. good practice. Quite all meta-analyses modulated their conclusions by addressing the need for large randomized controlled studies. Such a need seems now quite reinforced since results were until now expressed as pregnancy rate per cycle or pregnancy rate per couple, whereas single live birth rate and drop out rate are claimed to be taken into account nowadays. Moreover the level of controlled hyperstimulation is highly questionable with both hyperstimulation ovary syndrome and multiple pregnancy risks. Patients facing with failed i.u.i. cycles may turn to i.v.f. or i.c.s.i.. Interestingly data coming from the French national register (FIVNAT) did not show major differences between couples turning to i.v.f. (i.c.s.i.) after previously failed i.u.i. cycles or using directly i.v.f. (i.c.s.i.). Moreover the prognostic as evaluated on pregnancy rate per cycle was unchanged between such patients, which does not support some selection of patients by i.u.i. failure. Thus, although i.u.i. seems likely a cost-effective treatment in infertile couples, the precise conditions of its management (spontaneous or stimulated cycle, mono-, pauci- or multi-follicular induction) remain to be assessed. Indeed large controlled randomized studies including untreated group are required even if such design might have a non negligible cost. However these rather common treatments do have a high cost and any effort to rationalise them will have some economical impact. Another practical approach, although less ambitious, might consist in developing a per cycle registry which should allow to precise the French practice at a large national level.