胚胎移植后，必须一直卧床吗？

Protocol and registration

The study adhered to the principles of the preferred reporting guidelines for systematic reviews and meta-analyzes (PRISMA) (Moher et al., 2009a;b) (Figure 1). The study protocol was registered at http://www.crd.york.ac.uk/PROSPERO/CRD42020188716 before it was started. This study did not require institutional review board approval, since it is a systematic review.

〔此处图表、公式、附件或独立引文未收录，请核对期刊原文。〕

Eligibility criteria

The patient, intervention, comparison and outcome(s) (PICO) model was used to select the study population (Schardt et al., 2007). Women/couples who underwent IVF cycles were included, and the outcomes of patients prescribed bed rest vs. early ambulation were compared.

Search strategy

The authors systematically searched the literature for studies that matched the clinical question on MEDLINE, ClinicalTrials.gov, PubMed, and the Cochrane Library. A combination of Medical Subject Heading (MeSH) terms and/or relevant text words was performed to compare the clinical outcomes of rest vs. no post-transfer rest in IVF. The authors also checked citations on the Web of Science and manually searched the references of the articles. The searches were coordinated in May 2020. Updates to the search were made in June 2020. All records were evaluated for eligibility by two independent reviewers (J.R-P./E.C-B). All original peer-reviewed articles were included, regardless of study design.

The searches in electronic databases included the following combined search terms: (“Fertilization in Vitro” OR “in vitro fertilization” OR IVF OR “Reproductive Techniques, Assisted” OR “Oocyte Retrieval” OR “Egg Collection”) AND (“Bed Rest” OR “rest” OR “Bed Rest/adverse effects*” OR “Bed Rest/statistics & numerical data” OR “Postoperative period” OR “Early Ambulation) AND (“Embryo Transfer” OR “Embryo Transfer, Methods” OR “Embryo Implantation”) AND (“Live Birth Rate” OR “Live Birth” OR “Pregnancy Outcome” OR “Pregnancy” OR “Treatment Outcome” OR “Infant, Newborn” OR “Birth Rate” OR “Live-Birth”).

Study selection

Two authors (J.R-P./E.C-B) independently assessed article titles and abstracts. Duplicates were removed using the Zotero software and manually. The final decision to include/exclude articles was made after they were examined in their entirety. Discrepancies were resolved by discussion and consensus among the authors, with the participation of another author (MJ.G-C). To acquire the highest level of evidence, the authors selected randomized trials and observational studies, including cohort and case-control studies. Trials published only as abstracts, letters to the editor, editorials, or studies withdrawn from the literature after publication were excluded. Systematic reviews were also excluded, but references were checked first.

Intervention studies were eligible if: 1) they were randomized controlled trials (RCT), not RCT, prospective, retrospective observational studies, or cohort studies; 2) evaluated rest vs. no rest; 3) reported the LBR or clinical pregnancy rates (CPR) as outcome measures.

Data extraction and quality assessment

If the records were eligible, two reviewers (J.R-P. And E.C-B) collected and imported the data into an electronic database. The following data were captured: year of publication, study design, study period, intervention, number of patients (total, intervention, and comparator), age, rest time, biochemical pregnancy rate (BPR), CPR, LBR, and miscarriage rate (MR). Furthermore, for intervention studies, allocation concealment and blinding were also recorded.

The primary endpoint of the present meta-analysis was LBR. Although it could be argued that the implantation rate might be a better indicator of the effect of bed rest due to its temporality, the authors considered that the optimal clinical objective of an IVF cycle is the birth of a healthy infant that is ultimately sent home with their parents. Furthermore, the authors considered that the decision to prescribe rest or early ambulation might potentially impact the probability of miscarriage. As secondary endpoints, CPR and MR were studied. A LB was defined as any event of LB of a live product of conception after 24 weeks of gestation. LBR was defined as the proportion of LBs in relation to included patients. A clinical pregnancy was defined as the visualization of a gestational sac with a heartbeat after embryo transfer. CPR is the proportion of clinical pregnancies in relation to included patients. A biochemical pregnancy was defined as a blood pregnancy test >5mIUI/mL. The BPR is the proportion of biochemical pregnancies in relation to included patients. A miscarriage was defined as the loss of pregnancy before the 20^th week of gestation. MR was defined as the proportion of abortions in relation to the number of confirmed pregnancies.

Appraisal of certainty of evidence

Two authors (J.R-P./E.C-B.) Independently assessed the risk of bias of each study and the methodological quality of the included studies using the criteria described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins & Thompson, 2002). In addition, seven specific domains related to risk of bias were assessed: random sequence generation, allocation concealment, blinding of participants and staff, blinding of outcome assessment, incomplete results, selective data reporting, and other biases. Authors’ judgments were expressed as “low”, “high” or “unclear” risk of bias. The evaluation was carried out independently by two reviewers (J.R-P., E.C-B) and disagreement were resolved by discussion between the two parties including a third reviewer (MJ.G-C.).

Statistical analysis

Clinical outcomes were collected as dichotomous data. The results of the studies were combined in the meta-analysis using a Mantel-Haenszel fixed effect model for a pooled odds ratio (OR) and a 95% confidence interval (CI). Forest plots and the I² statistic were calculated for each study outcome and each group as a way to quantify the statistical heterogeneity of the included studies (Eden et al., 2011). I² was defined as: 0% to 40%: may not be important; 30% to 60%: may represent moderate heterogeneity; 50% to 90%: may represent substantial heterogeneity; 75% to 100%: considerable heterogeneity (Higgins & Thompson, 2002; Higgins et al., 2011). All calculations were performed with RevMan 5.4 (Review Manager, version 5.4, The Cochrane Collaboration, 2020).

Study selection

The search retrieved 27 citations, of which 14 were deemed eligible for review and four accepted for inclusion (Sharif et al., 1998; Bar-Hava et al., 2005; Purcell et al., 2007; Gaikwad et al., 2013) (Figure 1). Data from a total of 21,598 patients/cycles (rest: 20,138; early ambulation: 1,460) were included.

Description of included studies

Of the included studies, one was from the UK (Sharif et al., 1998), one from Israel (Bar-Hava et al., 2005), one from the USA (Purcell et al., 2007), and one from Spain (Gaikwad et al., 2013). There were two randomized controlled trials (Purcell et al., 2007; Gaikwad et al., 2013), one prospective (Bar-Hava et al., 2005), and one retrospective study (Sharif et al., 1998). Table 1 shows the description of the studies.

〔此处图表、公式、附件或独立引文未收录，请核对期刊原文。〕

ET: embryo transfer; n: sample size; RCT: randomized controlled study; NA: no information; IM: intramuscular; PO: orally

Assessment of the risk of study bias

According to the guidelines suggested by the Cochrane Collaboration, the quality of the included studies was moderate (Figure 2).

〔此处图表、公式、附件或独立引文未收录，请核对期刊原文。〕

Live birth rate

Two studies reported LBR (Purcell et al., 2007; Gaikwad et al., 2013). Patients prescribed bed rest after ET had an LBR of 43.6% vs. 52.5% in patients prescribed early ambulation (Table 2). The meta-analysis of the two studies yielded an OR 0.77 ((95% CI 0.5-1.2), I²=29%), which implies a 23% lower chance of having a LB among patients prescribed bed rest, although the difference was not statistically significant (Figure 3).

〔此处图表、公式、附件或独立引文未收录，请核对期刊原文。〕

〔此处图表、公式、附件或独立引文未收录，请核对期刊原文。〕

Clinical pregnancy rate

Three studies reported CPR (Sharif et al., 1998; Amarin & Obeidat, 2004; Purcell et al., 2007). Patients prescribed bed rest had a CPR of 18.7% vs. 25.3% in subjects prescribed early ambulation (Table 2). The meta-analysis of these three studies yielded an OR of 0.75 ((95% CI 0.7-0.9), I²=0%), which implies a 25% lower chance of clinical pregnancy among patients prescribed post-transfer rest (p<0.0001) (Figure 3).

Biochemical pregnancy rate

Three studies reported BPR (Sharif et al., 1998; Bar-Hava et al., 2005; Gaikwad et al., 2013). A BPR of 19.2% was observed in patients prescribed bed rest vs. 32.9% in individuals not prescribed bed rest (Table 2). The meta-analysis of these studies yielded an OR of 0.58 ((95% CI 0.5-0.7), I²=42%), which implies a 42% lower probability of biochemical pregnancy among patients prescribed bed rest (p<0.0001) (Figure 4).

〔此处图表、公式、附件或独立引文未收录，请核对期刊原文。〕

Miscarriage rate

Two studies reported MR (Purcell et al., 2007; Gaikwad et al., 2013), which, in patients prescribed bed rest, was 27.8% vs. 19.1% in subjects prescribed early ambulation (Table 2). The meta-analysis of these studies yielded an OR of 1.9 ((95% CI 1.0-3.4), I²=0%), which implies patients prescribed post-transfer rest are 1.9 times more likely to have an abortion (p<0.0001) (Figure 4).