The authors concluded that, based on the available evidence, continuing progesterone, beyond the first positive human chorionic gonadotropin test, after in vitro fertilisation (IVF) or intracytoplasmic sperm injection, might be unnecessary, but large randomised controlled trials were needed. These cautious conclusions reflect the limitations of the data and are likely to be reliable.

To assess the effects of early ending of progesterone treatment, on pregnancy outcomes, in women undergoing in vitro fertilisation (IVF) or intracytoplasmic sperm injection (ICSI).

The following databases were searched for articles included to July 2012, without language restrictions: MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials (CENTRAL), CBM, and Wanfang Data. Search terms were reported. Reference lists of relevant publications were screened.

Randomised controlled trials (RCTs) were eligible for inclusion if they compared early cessation of progesterone versus progesterone until the end of the first trimester, for pregnant women, after IVF or ICSI. The eligible outcomes were the live birth rate (babies born alive, after 24 weeks of gestation), the ongoing pregnancy rate (pregnancy beyond 12 weeks of gestation), and the miscarriage rate (the failure to achieve a live birth after a positive beta-human chorionic gonadotropin test).

Most of the included trials recruited patients with a positive beta-human chorionic gonadotropin test (on the 11th to 16th day after embryo transfer). The remaining trials recruited patients who were clinically pregnant (at five to seven weeks of gestation) or at the beginning of an IVF cycle. The trials used either agonistic or antagonistic gonadotropin-releasing hormone, as the controlled ovarian hyperstimulation protocol. The type and dose of progesterone, the time of administration, and the duration of treatment, varied across trials. Trials were conducted in Spain, Belgium, USA, Egypt, Denmark, or Germany.

The authors did not state how many reviewers assessed trials for inclusion.

The quality of the trials was assessed using the Cochrane risk of bias tool. Two reviewers independently assessed quality, with any disagreements resolved by consensus with a third reviewer.

The data were extracted on the event rates to calculate relative risks, with 95% confidence intervals. Where necessary, trial authors were contacted for additional relevant data.

Two reviewers independently extracted the data, with any disagreements resolved by consensus with a third reviewer.

The trial data were combined in a meta-analysis. The pooled relative risks, with 95% confidence intervals, were calculated using a random-effects model, where there was significant heterogeneity, or a fixed-effect model otherwise. Statistical heterogeneity was assessed using Χ² and Ι².

Sensitivity analyses were performed by excluding the trials with a quasi-randomised approach for patient allocation. A range of subgroup analyses was performed, including some based on the timing of randomisation, the timing of initiation of progesterone, the gonadotropin-releasing hormone (GnRH) analogue used for luteinising hormone surge inhibition, and the type and dose of progesterone.

Six RCTs were included in the review (1,201 patients). Randomisation and allocation concealment were adequate in four trials; one trial used a quasi-randomisation method for allocation. No trial blinded the investigators, participants or outcome assessors, to allocation.

There were no statistically significant differences between patients who ended progesterone early and those who continued on progesterone, for luteal-phase support, in their live birth rate (RR 0.95, 95% CI 0.86 to 1.05; two RCTs), miscarriage rate (RR 1.01, 95% CI 0.74 to 1.38; six RCTs), and ongoing pregnancy rate (RR 0.97, 95% CI 0.90 to 1.05; six RCTs). No significant heterogeneity was observed, except for the ongoing pregnancy rate (Ι²=73%).

Sensitivity analysis did not significantly alter the results. The results for the subgroup analyses were reported.

The available evidence suggested that continuing progesterone, beyond the first positive human chorionic gonadotropin test, after IVF or ICSI, might be unnecessary, but large randomised controlled trials were needed.

The review question was clear and was supported by appropriate inclusion criteria. A range of relevant databases was searched. No specific attempts were made to find unpublished trials, increasing the potential for publication bias. As the authors acknowledged, there were too few trials for a funnel plot to meaningfully assess publication bias. No language restrictions were applied, in the search, reducing the potential for language bias.

Attempts were made to minimise error and bias in data extraction and quality assessment, but it was unclear whether study selection was performed in by two people independently. Appropriate criteria were used to assess trial quality, and the analysis took into account some aspects of quality. Statistical heterogeneity was assessed, and appropriate methods were used to pool the trial results.

The authors' cautious conclusions reflect the limitations of the data and are likely to be reliable.

Practice: The authors did not state any implications for practice.

Research: The authors stated that further well-designed RCTs were required to investigate the optimal duration of progesterone, during early pregnancy, for women undergoing IVF or ICSI. These trials should recruit patients with early bleeding, advanced age, or polycystic ovary syndrome, as well as those with an inadequate human chorionic gonadotropin rise or endometriosis, to assess the outcomes for these patient subgroups.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.