We searched the MEDLINE, EMBASE, PubMed, Web of Science, Cochrane databases of systematic reviews, CINAHL, SCOPUS, China National Knowledge Infrastructure (CNKI), Wan Fang Database, and Chinese Science and Technology Periodical Database (VIP) for all relevant studies from the earliest data available to December 2022. The search and reporting procedure followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 checklist [31]. In addition, we manually retrieved and evaluated the reference lists of all the identified studies. All contents and methods were approved by the ethics committee of West China Hospital, Sichuan University (2021, Review No. 591).

Two independent researchers (JYW and JC) performed all searches. Our search strategy was based on the medical subject headings (Mesh) and free-text words, and the main Mesh were as follows: ‘Cardiac Catheters’, ‘Femoral Artery’, ‘early mobilization’, ‘Walking’ and ‘Bed Rest’. Two reviewers (JYW and SYT) independently screened and identified the studies for potential eligibility. We consulted the corresponding author to get a consensus about any controversy. The complete search strategy is shown in Supplementary Material.

The PICOS framework (P for the population of interest, I for intervention, C for comparison group, O for outcome, and S for study design) guided the study process [32, 33].

Studies that met the following inclusion criteria were included in this meta-analysis: (1) heart catheterization via femoral artery approach, (2) human studies comparing the safety of different lengths of bed rest with or without altering the patient’s position or the angle of beds, (3) participants older than 18 years of age, (4) randomized or quasi-randomized controlled trials involving more than ten patients in each group, (5) the studies provided no less than one clinical outcome, including bleeding, hematoma, pseudoaneurysm at the puncture site, back pain, urinary retention, or bladder catheterization, (6) language: English and Chinese. There were no restrictions concerning patient characteristics or healthcare settings.

We excluded studies if (1) any vascular closure device (VCD) or coagulants was applied to achieve the puncture site hemostasis except a bandage, sandbag, or manual compression, (2) the full text was not available, (3) only the position of the patients or the angle of the beds was altered, but the patients did not get out of bed, (4) the effects to combine early mobilization with other intervention variables (e.g., encouraging exercise and applying ice packs), (5) there was no specific ambulation timing or measurable outcome.

Shortening the post-cardiac catheterization duration of bed rest was the early mobilization group, and the longer duration of bed rest was regarded as the late ambulation group. According to the characteristics of the included studies, we classified different bed rest duration into the following four comparison subgroups: Group A: Comparing 2 h versus 4 h~6 h of bed rest; Group B: Comparing 3 h~4 h versus 5 h~6 h of bed rest; Group C: Comparing 4 h~6 h versus greater than or equal to 8 h of bed rest; Group D: Comparing 12 h versus 24 h of bed rest. The primary outcomes of interest were the incidence of bleeding and hematoma at the puncture site. The secondary outcomes were the incidence of back pain, pseudoaneurysm, urinary retention, and pain at the puncture site.

Using Excel, two authors (JYW and SYT) independently extracted and coded data from the qualified studies into standard tables. The original authors would be contacted for further information if data from the included studies were insufficient. Any disagreement was verified by the corresponding author. The items included: authors, year of publication, country, setting and location of the study, number of patients, the mean age of patients, study design, duration of the bed rest, hemostasis method, outcomes, and other relevant information.

The risk of bias in the included studies was independently evaluated by two authors (JYW and SYT) using the RoB 2 tool [34]. This tool assesses the following key areas of potential bias: randomization methods; deviations from intended intervention; missing outcome data; measurement of outcome; and selection of reported results. The judgment for the domain of RoB 2 is generated by an algorithm and can be ranked as low, high, or show some concerns. Each study is also given an overall judgment of RoB 2 based on the same options [34]. The corresponding author was available for arbitration in any disagreement regarding the ranking process and results.

This study was statistically analyzed using the Comprehensive Meta-Analysis software, Version 2 (Biostat, Englewood, NJ, USA). We pooled the odd ratio (OR), mean difference (MD), and 95% confidence intervals (CI) from separate studies to assess the intended effect sizes. Cochran’s Q test and the degree of inconsistency (I${}^2$ index) were used to evaluate the heterogeneity across studies. There was no heterogeneity if the I${}^2$ statistic was less than 25%, a low heterogeneity if the I${}^2$ statistic was 25–50%, and a moderate heterogeneity if the I${}^2$ statistic was 50–75%, while the I${}^2$ $>$ 75% reflected a high heterogeneity [35]. Funnel plots, Egger’s test [36], and Begg’s test [37] were used to analyze publication bias. Random-effect models that reflected the differences between each study were applied because of the heterogeneity across studies [38]. Subgroup analyses were performed based on the different bed rest duration. A p-value 0.05 was considered statistically significant.

The initial literature search yielded 14,443 studies, of which 1975 studies were excluded after removing duplicates. There were 94 studies eligible for further evaluation after 12,374 studies were excluded because of irrelevant titles and abstracts. At the full-text screening stage, only 80 studies were reviewed because 14 studies did not have the full text, and six studies met the inclusion and exclusion criteria. Then, we added eight eligible studies from the reference review process. Finally, 14 studies with 2653 participants were selected for this meta-analysis. A summary of the PRISMA flow diagram of the study selection process is shown in Fig. 1.

Fig. 1.

Flow diagram of the systematic searching process.

The 14 randomized controlled trials involving 2653 participants included 2 Chinese and 12 English studies. Two [39, 40] of the fourteen included studies were assigned to Group A with 325 participants, five studies [23, 41, 42, 43, 44] belonged to Group B with 1388 participants, five [26, 29, 45, 46, 47] were allocated to Group C with 705 participants, and the remaining two Chinese studies [48, 49] were in Group D with 235 participants. The mean age of the participants was between 53–67 years. These researches took place across three continents, five studies in North America, seven in Asia, and two in South America, of which four were conducted in China and four in the USA. All studies except for one conducted by Gu et al. [48] reported two arms. Based on our subgroup rules, we only selectively extracted the data from two groups (ambulation after 12 h versus 24 h). Six studies [23, 29, 42, 45, 46, 48] reported the indication for the cardiac catheterization, three for diagnosis [23, 42, 45] and three for therapy [29, 46, 48], respectively. Nine studies [23, 29, 40, 41, 42, 43, 46, 47, 48] used the manual compression method to obtain hemostasis, and a few studies employed additional methods such as bandage dressing or sandbags. Regarding heparin and sheath size, the usage regimen varied among the studies with some unavailable data. The characteristics are shown in Table 1 (Ref. [23, 26, 29, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49]).

Two authors (JYW and SYT) independently judged the risk of bias. Overall, ten of the fourteen studies were ranked as “high risk” and only one randomized controlled trial (RCT) was rated as “low risk”. Matte et al. [42] reported the research design and process according to all the domains of RoB 2, so we regarded this study as “low risk”. Three RCTs were judged as having some concerns considering overall risk [23, 40, 41]. Patients were doomed to be instructed about resting duration due to the nature of the study design; therefore, allocation concealment could not be achieved. We did not perceive any studies as having a high blinding risk of bias for participants or the individuals delivering the interventions. All the studies reported that patients followed the bed rest instructions. But in the deviations from intended interventions domain, only one study was scored as “high risk” due to not conforming to the established hemostatic protocol [43]. In the missing outcome data domain, we rated one study by Moeini et al. [46] to be at “high risk” because of the vague reporting of the study results. Figs. 2,3 depict the assessment results.

Fig. 2.

Risk of bias (RoB 2) assessment plot for the included randomized controlled trial studies.

Fig. 3.

Weighted summary plot of all the types of bias from the included studies.

We adopted a random effects analysis to assess the effect sizes. The forest plot showed the pooled effects of the six outcomes (bleeding, hematoma, back pain, pseudoaneurysm, urinary retention, and pain at the puncture site). In summary, six RCTs with a sample size of 569 found that the back pain of patients was significantly reduced by early mobilization (MD = 0.634, 95% CI: 0.23–1.038; p = 0.002). Nevertheless, early mobilization did not have any significant effects on bleeding (OR = 1.305, 95% CI: 0.683–2.494; p = 0.42), hematoma (OR = 1.328, 95% CI: 0.838–2.105; p = 0.227), pseudoaneurysm (OR = 1.442, 95% CI: 0.226–9.192; p = 0.698), urinary retention (OR = 2.62, 95% CI: 0.734–9.351; p = 0.138), and pain at the puncture site (MD = –0.019, 95% CI: –0.299–0.26; p = 0.892) among patients who underwent cardiac catheterization. The overall results of the meta-analysis are shown in Fig. 4.

Fig. 4.

The results of the overall odds ratio or mean difference for early mobilization effect among different outcomes. (a) Bleeding. (b) Hematoma. (c) Back pain. (d) Pseudoaneurysm. (e) Urinary retention. (f) Pain at the puncture site. The horizontal lines denote the 95% CI, the Square (◼) shows the point estimate (the size of the square corresponds to its weight); the diamond shows (◆) the combined overall effects of ambulation at 95% CI. CI, confidence interval.

Subgroup analyses were not possible owing to the lack of different bed rest duration groups concerning the outcome of pseudoaneurysm and pain at the puncture site. We only conducted subgroup analyses based on four outcomes: bleeding, hematoma, back pain, and urinary retention. The moderator variable was bed rest duration, including group A (2 h versus 4 h~6 h), group B (3 h~4 h versus 5 h~6 h), group C (4 h~6 h versus $\ge$8 h), and group D (12 h versus 24 h). Some subgroup analyses were unavailable because of the limited studies regarding different bed rest duration for each outcome.

We found the effect of patients’ back pain were statistically significant when they were they had instruction for early mobilization in groups B (MD = 0.737, 95% CI: 0.431–1.043; p = 0.000) and D (OR = 5.504, 95% CI: 1.646–18.407; p = 0.006). The forest plot shows the patients’ urinary retention was significantly decreased by early mobilization in group D (OR = 5.707, 95% CI: 1.859–17.521; p = 0.002). In contrast, early mobilization in group C (OR = 1.492, 95% CI: 0.317–7.013; p = 0.612) did not have an effect on the patients’ back pain. As for outcomes of bleeding and hematoma, the results showed no statistical significance among all the bed rest duration groups. Fig. 5 shows the details of the subgroup analysis.

Fig. 5.

Subgroup analysis to evaluate the effect of early mobilization. (a) Bleeding. (b) Hematoma. (c) Back pain. (d) Urinary retention. Group A, 2 h versus 4 h~6 h; Group B, 3 h~4 h versus 5 h~6 h; Group C, 4 h~6 h versus $\ge$8 h; Group D, 12 h versus 24 h. The horizontal lines denote the 95% CI, the Square (◼) shows the point estimate (the size of the square corresponds to its weight); the diamond shows (◆) the combined overall effects of ambulation at 95% CI. CI, confidence interval.

The I${}^2$ index of the effects of early mobilization on bleeding (I${}^2$ = 0%), hematoma (I${}^2$ = 0%), pseudoaneurysm (I${}^2$ = 0%), and pain at the puncture site (I${}^2$ = 21.31%) did not reflect heterogeneity, whereas the results of back pain (I${}^2$ = 72.06%) and urinary retention (I${}^2$ = 73.99%) showed a moderate heterogeneity. The publication bias of included studies reporting bleeding, hematoma, and back pain could be estimated by funnel plots, as shown in Fig. 6. Random-effect models were used in the overall meta-analysis process due to insufficient studies regarding some outcomes. The p-value of Egger’s test and Begg’s test for the effect of bleeding (p-value = 0.831 for Egger’s test and p-value = 0.714 for Begg’s test), hematoma (p-value = 0.415 for Egger’s test and p-value = 0.788 for Begg’s test), back pain (p-value = 0.621 for Egger’s test and p-value = 0.851 for Begg’s test), pseudoaneurysm (p-value = 0.941 for Egger’s test and p-value = 0.602 for Begg’s test), and urinary retention (p-value = 0.476 for Egger’s test and p-value = 0.497 for Begg’s test) did not indicate significant publication bias. Still, the possibility of publication bias cannot be denied. As shown in Table 2.

Fig. 6.

Funnel plot to assess publication bias in the effects of early ambulation on different outcomes. (a) Bleeding. (b) Hematoma. (c) Back pain. Diagonal lines represent pseudo-95% confidence intervals. The y-axis represents the standard error (weight in the pooled analysis), the x-axis indicates the effect size, and the vertical line shows the calculated estimated effect of different outcomes.