P < 0.05.

OR, odds ratio; PI, prediction interval; Trtm effect, treatment effect; n.a., not applicable; RoB, risk of bias; ITT, intention to treat; AT, as treated; AT+ET, as treated analysis of those participants who underwent an embryo transfer.

Results of all analyses for the primary outcome ‘live birth’. The intention to treat analysis was also performed as a sensitivity analysis according to the level of risk of bias, meaning that in the ‘low&medium RoB’ analysis all studies with high RoB were excluded, and the ‘low RoB’ analysis both studies with high and medium RoB were excluded.

Adjusted for age (grand mean-centered), BMI (grand mean-centered), duration of infertility (grand mean-centered, in months), number of previous failed transfers, cause of infertility (unexplained=reference, male factor, other), and type of treatment (no treatment/embryo transfer=reference, fresh, frozen). Thus, the reference patient has average age, average BMI, average duration of infertility, and no previous failed transfers. Additionally, in adjusted models, the reference patient has a diagnosis of unexplained infertility and has undergone no fertility treatment.

The number of participants varies between 3250 and 3257 in each of the 45 imputation rounds, because for 57 participants it was imputed whether or not an embryo transfer was performed.

AT+ET sensitivity analysis: one additional adjustment factor was added to the analysis, namely the number of embryos that were transferred in the study cycle. This was defined as a single embryo transfer (SET), double embryo transfer (DET), or triple or more embryo transfer (TET or more).

The outcome live birth was imputed for: ITT overall 100 participants, ITT low risk&some concerns 94 participants, ITT low risk 91 participants, AT 100 participants, AT+ET 6 participants.