Gonadotrophins versus clomiphene citrate with or without IUI in women with normogonadotropic anovulation and clomiphene failure: a cost-effectiveness analysis

STUDY QUESTION: Are six cycles of ovulation induction with gonadotrophins more cost-effective than six cycles of ovulation induction with clomiphene citrate (CC) with or without IUI in normogonadotropic anovulatory women not pregnant after six ovulatory cycles with CC? SUMMARY ANSWER: Both gonadotrophins and IUI are more expensive when compared with CC and intercourse, and gonadotrophins are more effective than CC. WHAT IS KNOWN ALREADY: In women with normogonadotropic anovulation who ovulate but do not conceive after six cycles with CC, medication is usually switched to gonadotrophins, with or without IUI. The cost-effectiveness of these changes in policy is unknown. STUDY DESIGN, SIZE, DURATION: We performed an economic evaluation of ovulation induction with gonadotrophins compared with CC with or without IUI in a two-by-two factorial multicentre randomized controlled trial in normogonadotropic anovulatory women not pregnant after six ovulatory cycles with CC. Between December 2008 and December 2015 women were allocated to six cycles with gonadotrophins plus IUI, six cycles with gonadotrophins plus intercourse, six cycles with CC plus IUI or six cycles with CC plus intercourse. The primary outcome was conception leading to a live birth achieved within 8 months of randomization. PARTICIPANTS/MATERIALS, SETTING, METHODS: We performed a cost-effectiveness analysis on direct medical costs. We calculated the direct medical costs of ovulation induction with gonadotrophins versus CC and of IUI versus intercourse in six subsequent cycles. We included costs of medication, cycle monitoring, interventions, and pregnancy leading to live birth. Resource use was collected from the case report forms and unit costs were derived from various sources. We calculated incremental cost-effectiveness ratios (ICER) for gonadotrophins compared to CC and for IUI compared to intercourse. We used non-parametric bootstrap resampling to investigate the effect of uncertainty in our estimates. The analysis was performed according to the intention-to-treat principle. MAIN RESULTS AND THE ROLE OF CHANCE: We allocated 666 women in total to gonadotrophins and IUI (n = 166), gonadotrophins and intercourse (n = 165), CC and IUI (n = 163), or CC and intercourse (n = 172). Mean direct medical costs per woman receiving gonadotrophins or CC were €4495 versus €3006 (cost difference of €1475 (95% CI: €1457-€1493)). Live birth rates were 52% in women allocated to gonadotrophins and 41% in those allocated to CC (relative risk (RR) 1.24:95% CI: 1.05-1.46). The ICER was €15 258 (95% CI: €8721 to €63 654) per additional live birth with gonadotrophins. Mean direct medical costs per woman allocated to IUI or intercourse were €4497 versus €3005 (cost difference of €1510 (95% CI: €1492-€1529)). Live birth rates were 49% in women allocated to IUI and 43% in those allocated to intercourse (RR = 1.14:95% CI: 0.97-1.35). The ICER was €24 361 (95% CI: €-11 290 to €85 172) per additional live birth with IUI. LIMITATIONS, REASONS FOR CAUTION: We allowed participating hospitals to use their local protocols for ovulation induction and IUI, which may have led to variation in costs, but which increases generalizability. Indirect costs generated by transportation or productivity loss were not included. We did not evaluate letrozole, which is potentially more effective than CC. WIDER IMPLICATIONS OF THE FINDINGS: Gonadotrophins are more effective, but more expensive than CC, therefore, the use of gonadotrophins in women with normogonadotropic anovulation who have not conceived after six ovulatory CC cycles depends on society's willingness to pay for an additional child. In view of the uncertainty around the cost-effectiveness estimate of IUI, these data are not sufficient to make recommendations on the use of IUI in these women. In countries where ovulation induction regimens are reimbursed, policy makers and health care professionals may use our results in their guidelines. STUDY FUNDING/COMPETING INTEREST(S): This trial was funded by the Netherlands Organization for Health Research and Development (ZonMw number: 80-82310-97-12067). The Eudract number for this trial is 2008-006171-73. The Sponsor's Protocol Code Number is P08-40. CBLA reports unrestricted grant support from Merck and Ferring. BWM is supported by a NHMRC Practitioner Fellowship (GNT1082548) and reports consultancy for Merck, ObsEva and Guerbet.NTR1449