A novel approach for retrospectively estimating the efficiency of PGT-A testing

Objective: Prior studies have estimated the efficiency of PGT-A using published implantation and aneuploidy rates [1]. Here we developed a more rigorous approach that incorporates a novel methodology for imputing the likelihood of failed untested transfers resulting from aneuploidy.

Materials and Methods: Historical de-identified electronic medical records were collected from a single IVF clinic in the United States between 2015-2020. Dataset 1 consisted of PGT-A results for 3,680 blastocysts from 773 patients with patient age and blastocyst grades. Dataset 2 consisted of 1,449 autologous, single-blastocyst FET cycles with patient age, blastocyst grade, and transfer outcomes. Using dataset 1, we developed a mixed effects logistic regression model to predict the likelihood of a blastocyst having an abnormal test result using age, biopsy day, and grade. Using dataset 2, we calculated the observed non-PGT-A implantation rates and PGT-A implantation rates stratified by age. Then, for failed non-PGT-A transfers, we applied the mixed effects model to calculate the probability that a failed blastocyst was abnormal, and removed these embryos (at a rate proportional to their predicted probability) to calculate an idealized PGT-A implantation rate. Finally, the idealized and observed PGT-A implantation rates were compared to estimate the loss caused by PGT-A testing.

Results: Per-transfer non-PGT-A implantation rates were 30%-57%, while PGT-A implantation rates were 60%-68%. After imputing the likelihood of failed untested transfers resulting from aneuploidy, the idealized PGT-A implantation rates were 64%-71%. The estimated implantation loss due to PGT-A testing was 1% for ages <=33, 0% for ages 34-36, 16% for ages 37-39, and 3% for ages >=40.

Conclusions: We estimated that implantation losses due to PGT-A testing for this clinic are near 0% for younger patient groups, and approximately 3%-16% for older patient groups. While it is not possible to determine the cause of potential implantation loss, it could result from harm from the biopsy procedure or discarding of viable embryos.

Impact Statement: PGT-A efficiency appears to be high, with near-zero implantation losses for younger patients and approximately 3% to 16% implantation losses for older patients.

References: [1] Paulson RJ. Hidden in plain sight: the overstated benefits and underestimated losses of potential implantations associated with advertised PGT-A success rates. Hum Reprod. 2020;35(3):490-493.