BackgroundIn the last two decades, pregnancy rates for patients undergoing in-vitro fertilisation (IVF) have significantly increased. Some of the major advances responsible for this improvement were the introduction of controlled ovarian stimulation (COS) for the induction of multiple follicle development, and the utilisation of mid-luteal gonadotropin-releasing hormone agonists to achieve pituitary down-regulation and full control of the cycle. As a result, a combination of a gonadotropin-releasing hormone agonist with high doses (150-450 IU/day) of recombinant follicle-stimulating hormone has become the current standard approach for ovarian stimulation. However, given the heterogeneity of patients embarking on IVF, and the fact that many different drugs can be used alone or in different combinations (generating multiple potential protocols of controlled ovarian stimulation), we consider the need to identify special populations of patients and adapt treatment protocols accordingly, and to implement a more individualised approach to COS.DiscussionStudies on mild, minimal and natural IVF cycles have yielded promising results, but have focused on fresh embryo transfers and included relatively young patient populations who generally have the potential for more favourable outcomes. The efficacy of these protocols in patients with a poorer prognosis remains to be tested. When comparing protocols for COS, it is important to think beyond current primary endpoints, and to consider the ideal quality and quantity of oocytes and embryos being produced per stimulated patient, in order to achieve a pregnancy. We should also focus on the cumulative pregnancy rate, which is based on outcomes from fresh and frozen embryos from the same cycle of stimulation. Individualised COS (iCOS) determined by the use of biomarkers to test ovarian reserve has the potential to optimise outcomes and reduce safety issues by adapting treatment protocols according to each patient's specific characteristics. As new objective endocrine, paracrine, functional and/or genetic biomarkers of response are developed, iCOS can be refined further still, and this will be a significant step towards a personalised approach for IVF.ConclusionsA variety of COS protocols have been adopted, with mixed success, but no single approach is appropriate for all patients within a given population. We suggest that treatment protocols should be adapted for individual patients through iCOS; this approach promises to be one of the first steps towards implementing personalised medicine in reproductive science.