Reduced live-birth rates after IVF/ICSI in women with previous unilateral oophorectomy: results of a multicentre cohort study.

Tekla Lind, Jan Holte, Jan I Olofsson, Nermin Hadziosmanovic, Johannes Gudmundsson, Elizabeth Nedstrand, Mikael Lood, Lars Berglund, Kenny Rodriguez-Wallberg

Is there a reduced live-birth rate (LBR) after IVF/ICSI treatment in women with a previous unilateral oophorectomy (UO)? A significantly reduced LBR after IVF/ICSI was found in women with previous UO when compared with women with intact ovaries in this large multicentre cohort, both crudely and after adjustment for age, BMI, fertility centre and calendar period and regardless of whether the analysis was based on transfer of embryos in the fresh cycle only or on cumulative results including transfers using frozen-thawed embryos. Similar pregnancy rates after IVF/ICSI have been previously reported in case-control studies and small cohort studies of women with previous UO versus women without ovarian surgery. In all previous studies multiple embryos were transferred. No study has previously evaluated LBR in a large cohort of women with a history of UO. This research was a multicentre cohort study, including five reproductive medicine centres in Sweden: Carl von Linné Clinic (A), Karolinska University Hospital (B), Uppsala University Hospital (C), Linköping University Hospital (D) and Örebro University Hospital (E). The women underwent IVF/ICSI between January 1999 and November 2015. Single embryo transfer (SET) was performed in approximately 70% of all treatments, without any significant difference between UO exposed women versus controls (68% versus 71%), respectively (P = 0.32), and a maximum of two embryos were transferred in the remaining cases. The dataset included all consecutive treatments and fresh and frozen-thawed cycles. The exposed cohort included 154 women with UO who underwent 301 IVF/ICSI cycles and the unexposed control cohort consisted of 22 693 women who underwent 41 545 IVF/ICSI cycles. Overall, at the five centres (A-E), the exposed cohort underwent 151, 34, 35, 41 and 40 treatments, respectively, and they were compared with controls of the same centre (18 484, 8371, 5575, 4670 and 4445, respectively). The primary outcome was LBR, which was analysed per started cycle, per ovum pick-up (OPU) and per embryo transfer (ET). Secondary outcomes included the numbers of oocytes retrieved and supernumerary embryos obtained, the Ovarian Sensitivity Index (OSI), embryo quality scores and cumulative pregnancy rates. We used a Generalized Estimating Equation (GEE) model for statistical analysis in order to account for repeated treatments. The exposed (UO) and control women's groups were comparable with regard to age and performance of IVF or ICSI. Significant differences in LBR, both crude and age-adjusted, were observed between the UO and control groups: LBR per started cycle (18.6% versus 25.4%, P = 0.007 and P = 0.014, respectively), LBR/OPU (20.3% versus 27.1%, P = 0.012 and P = 0.015, respectively) and LBR/ET (23.0% versus 29.7%, P = 0.022 and P = 0.025, respectively). The differences in LBR remained significant after inclusion of both fresh and frozen-thawed transfers (both crude and age-adjusted data): LBR/OPU (26.1% versus 34.4%, P = 0.005 and P = 0.006, respectively) and LBR/ET (28.3% versus 37.1%, P = 0.006 and P = 0.006, respectively). The crude cancellation rate was significantly higher among women with a history of UO than in controls (18.9% versus 14.5%, P = 0.034 and age-adjusted, P = 0.178). In a multivariate GEE model, the cumulative odds ratios for LBR (fresh and frozen-thawed)/OPU (OR 0.70, 95% CI 0.52-0.94, P = 0.016) and LBR (fresh and frozen-thawed)/ET (OR 0.68, 95% CI 0.51-0.92, P = 0.012) were approximately 30% lower in the group of women with UO when adjusted for age, BMI, reproductive centre, calendar period and number of embryos transferred when appropriate. The OSI was significantly lower in women with a history of UO than in controls (3.6 versus 6.0) and the difference was significant for both crude and age-adjusted data (P = <0.001 for both). Significantly fewer oocytes were retrieved in treatments of women with UO than in controls (7.2 versus 9.9, P = <0.001, respectively). Due to the nature of the topic, this is a retrospective analysis, with all its inherent limitations. Furthermore, the cause for UO was not possible to obtain in all cases. A diagnosis of endometriosis was also more common in the UO group, i.e. a selection bias in terms of poorer patient characteristics in the UO group cannot be completely ruled out. However, adjustment for all known confounders did not affect the general results. To date, this is the largest cohort investigated and the first study indicating an association of achieving reduced live birth after IVF/ICSI in women with previous UO. These findings are novel and contradict the earlier notion that IVF/ICSI treatment is not affected, or is only marginally affected by previous UO. None. Not applicable.