P-703 Developing individualized treatment strategies using detailed subgroup analysis to optimise pregnancy outcomes for women 40 and above

Abstract

Abstract Study question What are the prognostic factors and treatment strategies for women 40 and above within different ovarian reserve subgroups? Summary answer Women in each ovarian reserve subgroup have different prognostic factors of pregnancy success and specific treatment strategies are needed for each subgroup. What is known already The number of women 40 and above seeking subfertility treatment is increasing and many are willing to undergo treatment despite extremely poor prognosis. POSEIDON criteria suggested the usage of ovarian reserve and female age as a basis for the study of the management of women in different prognostic categories. However, this classification may not be detailed enough to lead to specific recommendations for individualised management in women 40 and above, as there is rapid decline of oocyte quality and quantity. Large observational studies may lack finer detail due to challenges in recruiting women within different ovarian reserve subgroups. Study design, size, duration We carried out a detailed retrospective analysis of 382 IVF/ICSI cycles for 253 women from January 2019 to December 2021 in a special clinical and academic program for reproductive ageing and fertility. Inclusion criteria were all women who had IVF/ICSI cycles aged 40 and over. Analysis was done for all fresh cycles including cancelled cycles, failed fertilization, no eggs collected and embryo arrest. Cycles resulting in freezing of all embryos were excluded in analysis. Participants/materials, setting, methods The participants were divided into three main subroups: Group A, “normal ovarian reserve” (AMH 2.5-29.9pmol/l, n = 271 cycles); Group B, “extremely poor ovarian reserve” (AMH <2.5pmol/l, n = 85 cycles); and Group C “extremely high ovarian reserve” (PCOS or AMH ≥30.0pmol/l, n = 26 cycles). Chi-square and Mann-Whitney U tests were used to compare pregnant and non-pregnant groups for nominal and continuous data respectively. Multi-variate logistic regression was performed when indicated. All descriptive statistics are stated in median and percentages. Main results and the role of chance In Group A, univariate analysis showed that pregnancies were associated with lower female (41 vs 42) and male age (40 vs 43.5), administering long agonist versus antagonist protocol (73% vs 27%), the absence of androgen pretreatment (12% vs 88%), higher oestradiol level (7681pmol/l vs 6114pmol/l), higher number of oocytes retrieved (8 vs 4) and normally fertilized (4 vs 3), and embryo transfer on day 3 compared to 2 or 5 (all p < 0.05). The use of combined FSH/LH vs FSH only for ovarian stimulation given for 12 days at 300IU daily, higher endometrial thickness (11.3mm vs 10.6mm), IVF versus ICSI (65% vs 35%), and addition of subcutaneous progesterone for luteal support versus vaginal progesterone only (51% vs 43%) were trending towards but were not significant (p > 0.05). Multivariate logistic regression analysis revealed lower female and male age, less number of IVF cycles, and higher number of embryos transferred as independent factors contributing to pregnancy. In Group B, significant predictors of pregnancy included younger female age, higher gravidity, previous successful IVF and higher number of eggs fertilized (p < 0.05). In Group C, factors that contributed to positive pregnancy outcome included antagonist cycles versus agonist and usage of FSH and LH versus FSH alone (p > 0.05). Limitations, reasons for caution The retrospective nature of the study design limits control of potential confounding variables. Although there were clear trends, in a few sub-categories there were insufficient numbers to reach statistical significance. Continuing data collection is necessary to achieve adequate power for smaller groups especially Group B and C. Wider implications of the findings This detailed analysis will allow us to better understand the prognostic factors in women 40 and above, especially for women with extremes of ovarian reserve, who may benefit substantially from individualised care. We recommend standardisation of data collection and development of artificial intelligence algorithms to streamline management in the future. Trial registration number N/A