Randomized clinical trial of preconception low dose aspirin use and pregnancy loss: results from the eager trial

miscarriages, IVF/PGS (PGS) are increasingly performed in RPL patients with limited clinical evidence. Our objective is to determine if PGS is cost-effective compared to EM in achieving live birth (LB) and preventing clinical miscarriage (CM) for patients with unexplained RPL. DESIGN: A decision analysis of cost incurred and clinical outcomes with PGS versus EM. MATERIALS AND METHODS: A decision-analytic model was constructed using TreeAge Pro 2013. Probabilities for IVF/PGS were derived from published data of outcomes of IVF with 24-chromosome PGS involving 2,282 embryos and 287 PGS cycles followed by fresh embryo transfer (ET) in patients with unexplained RPL. Probabilities for EM were obtained from a prospective longitudinal study of 325 patients with unexplained RPL who underwent clinical evaluation and attempted spontaneous conception with close follow-up. Cost data was obtained from literature and adjusted to 2014 US dollars . For both groups, the average cost of pre-conception counseling and baseline RPL workup including parental karyotyping, maternal anti-phospholipid antibody testing and uterine cavity evaluation was $4,377 . No additional cost was incurred in the EM group. The average cost of IVF with one ET was $18,227 and the average cost of PGS was $4,268 . Cost to achieve live birth and cost to prevent one CM were the primary outcomes of the analysis. In the first model, all patients underwent one PGS cycle and ET, if a euploid embryo was produced. In the second model, patients did a second cycle if they did not conceive with the first. Base case, threshold, and one-way sensitivity analyses and a Monte Carlo simulation were performed to assess the robustness of the model. RESULTS: Cost effectiveness analysis favored EM over PGS. The CP rate per attempt is 35.6% with IVF/PGS compared to 69.5% with EM. Clinical outcomes, cost per LB, and cost to prevent one CM using one and two cycles of PGS compared to EM is shown in Table 1. The model was robust to one-way sensitivity analysis and Monte Carlo simulation.