Tremellen 2000.

Methods	Multicentre, parallel group study Randomisation stratified for centre, hence for fresh or frozen embryo transfer Number of women randomised: 600 (302 in the intervention group; 298 in the control group) Number of cancellations: 122 cycles (60 in the intervention group, 62 in the control group) which did not reach embryo transfer Number of cycles analysed: 478
Participants	Country of authors: Spain and Australia Inclusion criteria: women aged 18 to 40 years, in a stable relationship, undergoing ART with fresh (in the Spanish centre) or frozen (in the Australian centre) embryo transfer Exclusion criteria: hepatitis B, or C or HIV positivity in the male partner Mean age ± SD: Spanish centre: intervention group: 33.3 ± 3.3 years, control group: 33.2 ± 3.3 years; Australian centre: intervention group: 33.8 ± 4.4 years, control group: 33.1 ± 4.4 years Mean duration of infertility ± SD: Spanish centre: intervention group: 4.7 ± 2.6 years, control group: 4.1 ± 2.8 years; Australian centre: intervention group: 5 ± 2.5 years, control group: 5.1 ± 2.8 years Women with primary infertility: not reported Setting: assisted reproduction programmes in Spain (Valencia and Murcia Clinics of the Instituto Valenciano de Infertilidad) and Australia (University of Adelaide Reproductive Medicine Unit)
Interventions	Interventions: fresh ART cycles in Spain: vaginal intercourse at least twice, 12 hours before oocyte pick up and 12 hours after embryo transfer Frozen embryo transfer cycles in Australia: vaginal intercourse at least once, during the period between four days before and two days after embryo transfer Comparator: abstinence during the same period.
Outcomes	Biochemical pregnancy, clinical pregnancy at 6 to 8 weeks of gestation, multiple pregnancy
Notes	We contacted the corresponding author, who understandably did not have the missing data for this trial, which was published 18 years ago.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer generated block randomisation list stratified for centre
Allocation concealment (selection bias)	Low risk	Sealed opaque envelopes managed by 3rd party
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants were not blinded, though they were asked about adherence to the protocol
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Objectively measured outcome
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	25% of cycles cancelled after randomisation. However, the proportion of cancelled cycles were similar between the groups and ITT analysis was done.
Selective reporting (reporting bias)	Low risk	Clinical pregnancy reported
Other bias	High risk	Significantly more embryos transferred in the control group