Fig. 1.

a Fertilization rates with and without CT as predicted by the best supported mixed-effect logistic regression model which controlled for the confounding effect of age. The shaded areas indicate the 95% confidence intervals. b The probability for CT embryos being superior increased with an increasing proportion of oocytes with CT, as predicted by the best logistic regression model which contained proportions of oocytes with CT and a total dose of FSH used for stimulation. The curve was drawn while keeping a total dose of FSH at the median value. If embryos with and without CT are of equal quality, then the predicted curve should come exactly through the cross of the dashed lines which indicates a probability of 50% for a proportion of 0.5. As the curve intersects the proportion of 0.5 at the probability 0.723, the oocytes subjected to CT clearly result in more embryos of higher quality. c The probability for CT embryos of being superior decreases with the increasing total dose of FSH as predicted by the same model while keeping the proportion of CT oocytes at the median value. The probability, however, remains above 0.5 over the whole range of total FSH dose values