OBJECTIVE Which clinical and laboratory factors affect live birth rate (LBR) after frozen-thawed embryo transfer (FET)? CONCLUSIONS Top quality embryo characteristics, endometrial preparation protocol, number of embryos transferred and BMI affected independently the LBR in FET. BACKGROUND FET is an important part of present-day IVF/ICSI treatment. There is limited understanding of the factors affecting success rates after FET. METHODS This is a two-centre retrospective cohort study. Analysis was carried out on 1972 consecutive FET cycles in 1998-2007, with embryos frozen on Day 2. The primary outcome was LBR per cycle. METHODS We assessed the independent effect on LBR of the following variables: female age, female age at embryo freezing, BMI, diagnosis, primary versus secondary infertility, fertilization by IVF versus ICSI, pregnancy in the fresh cycle, type (spontaneous, spontaneous with luteal progesterone and estrogen/progesterone substitution) and rank of the FET cycle, as well as number and presence (yes versus no) of top quality embryo(s) at freezing, thawing and transfer, damaged thawed embryos and overnight culture. RESULTS In 78% of the cycles with top quality embryos frozen (n = 1319), at least one embryo still had high-quality morphology after thawing. Top quality embryo morphology observed at any stage of culture improved the outcome even if high-quality characteristics disappeared before transfer. LBRs after the transfer of a top quality embryo were similar in the FET (24.9%) and fresh cycles of the same period (21.9%). The chance of live birth increased significantly if ≥1 top quality embryo was present at freezing (odds ratio (OR) 1.85, 95% confidence interval (CI) 1.10-3.14), at thawing (OR 1.93, CI 1.20-3.11) or at transfer (OR 3.41, CI 2.12-5.48). Compared with spontaneous cycles with luteal support, purely spontaneous cycles (OR 0.58, CI 0.40-0.84) and hormonally substituted FET (OR 0.47, CI 0.32-0.69) diminished the odds of pregnancy. BMI (OR 0.96, CI 0.92-0.99) and transfer of two embryos versus one (OR 1.45, CI 1.08-1.94) were other factors that improved LBR after FET. CONCLUSIONS The sample sizes available in some subanalyses were small, limiting the power of the study. CONCLUSIONS The presence of ≥1 top quality embryo at any step of the freezing and thawing process increases the chance of pregnancy. The data do not support the freezing of all embryos for transfer in order to improve the outcome. A top quality embryo transferred in FET may even have the same potential as in a fresh cycle. On the contrary, LBR in the group with no top quality embryos frozen was quite low (10.4%), raising the question of whether a re-evaluation of freezing criteria is necessary to avoid costly treatments with a low success rate.