Timing of dexamethasone initiation during controlled ovarian stimulation and live birth after IVF/ICSI: a retrospective cohort study.

BACKGROUND

Dexamethasone (DXM) is used as an empirical adjuvant in IVF/ICSI, but evidence for its benefit remains inconsistent, and the clinical importance of initiation timing during controlled ovarian stimulation (COS) is unclear.

METHODS

We retrospectively analyzed 644 COS cycles at a tertiary reproductive center. Cycles were classified as no DXM (n = 185), early-follicular DXM (stimulation days 0-4; n = 296), mid-follicular DXM (days 5-7; n = 69), or late/trigger-day DXM (day ≥8 or hCG day; n = 94). DXM was prescribed at the treating physician's discretion, mainly for cycles considered at increased risk of premature progesterone elevation, and was administered as oral dexamethasone acetate once daily at 0.75-1.125 mg/day, predominantly 0.75 mg/day. Ovarian response, embryo development, and total usable embryos were analyzed in the full COS cohort. Pregnancy outcomes were analyzed among post-transfer cycles, and neonatal outcomes among deliveries. Multivariable logistic regression, propensity score-based inverse probability of treatment weighting, and mediation analysis were used to assess associations and address confounding.

RESULTS

Mid- and late/trigger DXM cycles generated more transferable and frozen embryos and more total usable embryos than cycles without DXM. In the post-transfer cohort, early- and mid-follicular DXM were associated with higher live-birth rates than no DXM, whereas late/trigger DXM showed no clear live-birth advantage. Adjusted and propensity score-weighted models produced consistent estimates. Mediation analysis did not show significant evidence that total usable embryos mediated the association between early DXM and live birth. Neonatal outcomes, including gestational age, birthweight, preterm birth, low birthweight, and congenital malformations, did not differ significantly by DXM timing.

CONCLUSIONS

In this retrospective cohort, DXM timing was associated with distinct reproductive outcomes. Early- and mid-follicular DXM were associated with higher live birth after transfer, whereas late/trigger DXM did not show clear clinical benefit. These findings should be interpreted cautiously given the clinically selected nature of DXM use.