Multivariate analysis of the factors associated with live births during in vitro fertilisation in Southeast Asia: a cross-sectional study of 104,015 in vitro fertilisation records in Taiwan

Purpose

To investigate the factors associated with live births and the interaction between age and the number of embryos transferred after in vitro fertilisation (IVF) treatment.

Methods

This study analyses data from a population-based-assisted reproductive database of all registered artificial reproduction institutions (n = 80) from 2010 to 2016 in Taiwan. The probability of a live birth in correlation with the treatment parameters was measured with multivariate logistic regression analyses using the generalised additive model (GAM) and Pearson’s chi-square exact test.

Results

A total of 104,015 IVF treatments performed between 2010 and 2016 were included in our analysis. From these treatments, 31,467 (30.3%) were successfully delivered, and 40,565 test-tube babies were born. Pearson’s chi-square exact test indicated that parents’ age, cause of infertility, type of ovarian stimulation, additional assisted reproductive technology techniques, donated egg or sperm, fresh or frozen embryo, presence or absence of ovarian hyperstimulation syndrome, and day of embryo transfer were significantly associated with live births after an IVF cycle (p < 0.05). Multiple logistic regression analysis with the GAM revealed that the odds of a live birth with IVF treatment in patients < 34 years of age were 2.55 times higher than that in patients ≥ 45 years of age (odds ratio = 2.55, 95% confidence interval = 1.69–2.90) for patients who underwent a single-embryo transfer (SET); a similar pattern was observed when two or more embryos were transferred. Egg donation, the assisted hatching technique, oral ovarian stimulation agents, and implantation of frozen embryos during SET were shown to improve the chance of a live birth by 29–90%. Implantation of the embryo after the 5th day of culture yielded the highest odds of a live birth. The interaction plot revealed that maternal age, especially < 40 years, was associated with the probability of a live birth. SET and double-embryo transfer showed similar associations with the probability of a live birth across age groups. Transferring more than two embryos might reduce the probability of a live birth during IVF treatment for women ≥ 40 years of age.

Conclusions

Implanting a greater number of embryos did not improve the age-related decrease in fertility for patients undergoing IVF. Therefore, we suggested that ≤ 2 blastocysts could be transferred during IVF treatments for women ≥ 40 years. Transferring a blastocyst on day 5 of culture was associated with a significant increase in the odds of a live birth resulting from IVF.