STUDY QUESTION: Does an increased FSH dose result in higher cumulative live birth rates in women with a predicted poor ovarian response, apparent from a low antral follicle count (AFC), scheduled for IVF or ICSI? SUMMARY ANSWER: In women with a predicted poor ovarian response (AFC < 11) undergoing IVF/ICSI, an increased FSH dose (225/ 450 IU/day) does not improve cumulative live birth rates as compared to a standard dose (150 IU/day). WHAT IS KNOWN ALREADY: In women scheduled for IVF/ICSI, an ovarian reserve test (ORT) can predict ovarian response to stimulation. The FSH starting dose is often adjusted based on the ORT from the belief that it will improve live birth rates. However, the existing RCTs on this topic, most of which show no benefit, are underpowered. STUDY DESIGN, SIZE, DURATION: Between May 2011 and May 2014, we performed an open-label multicentre RCT in women with an AFC < 11 (Dutch Trial Register NTR2657). The primary outcome was ongoing pregnancy achieved within 18 months after randomization and resulting in a live birth. We needed 300 women to assess whether an increased dose strategy would increase the cumulative live birth rate from 25 to 40% (two-sided alpha-error 0.05, power 80%). PARTICIPANTS/MATERIALS, SETTING, METHODS: Women with an AFC ≤ 7 were randomized to an FSH dose of 450 IU/day or 150 IU/day, and women with an AFC 8–10 were randomized to 225 IU or 150 IU/day. In the standard group, dose adjustment was allowed in subsequent cycles based on pre-specified criteria. Both effectiveness and cost-effectiveness of the strategies were evaluated from an intention-to-treat perspective. MAIN RESULTS AND THE ROLE OF CHANCE: In total, 511 women were randomized, 234 with an AFC ≤ 7 and 277 with an AFC 8–10. The cumulative live birth rate for increased versus standard dosing was 42.4% (106/250) versus 44.8% (117/261), respectively [relative risk (RR): 0.95 (95%CI, 0.78–1.15), P = 0.58]. As an increased dose strategy was more expensive [delta costs/woman: €1099 (95%CI, 562–1591)], standard FSH dosing was the dominant strategy in our economic analysis. LIMITATIONS, REASONS FOR CAUTION: Despite our training programme, the AFC might have suffered from inter-observer variation. As this open study permitted small dose adjustments between cycles, potential selective cancelling of cycles in women treated with 150 IU could have influenced the cumulative results. However, since first cycle live birth rates point in the same direction we consider it unlikely that the open design masked a potential benefit for the individualized strategy. WIDER IMPLICATIONS OF THE FINDINGS: Since an increased dose in women scheduled for IVF/ICSI with a predicted poor response (AFC < 11) does not improve live birth rates and is more expensive, we recommend using a standard dose of 150 IU/day in these women. STUDY FUNDING/COMPETING INTEREST(S): This study was funded by The Netherlands Organisation for Health Research and Development (ZonMW number 171102020). T.C.T., H.L.T. and S.C.O. received an unrestricted personal grant from Merck BV. H.R.V. receives monetary compensation as a member on an external advisory board for Ferring pharmaceutical BV. B.W.J.M. is supported by a NHMRC Practitioner Fellowship (GNT1082548) and reports consultancy for OvsEva, Merck and Guerbet. F.J.M.B. receives monetary compensation as a member of the external advisory board for Ferring pharmaceutics BV (the Netherlands) and Merck Serono (the Netherlands) for consultancy work for Gedeon Richter (Belgium) and Roche Diagnostics on automated AMH assay development (Switzerland) and for a research cooperation with Ansh Labs (USA). All other authors have nothing to declare.