Same-day Discharge Following Percutaneous Coronary Intervention: A Systematic Review and Meta-analysis of Randomised Controlled Trials

Percutaneous coronary intervention (PCI) is one of the most commonly performed medical procedures worldwide.¹ Currently, patients undergoing PCI are typically monitored during an overnight stay (ONS) prior to hospital discharge to facilitate timely detection and management of post-procedural complications, including MI, stent thrombosis, bleeding and renal impairment.^2,3 During this period, diagnostic evaluations (such as serial measurements of haemoglobin, troponin and creatinine levels, and vascular ultrasonography) are performed alongside therapeutic interventions, including IV fluid administration and, when indicated, repeat angiography.

However, in recent years, advances in PCI techniques, stent technology and antithrombotic strategies have significantly reduced the risk of adverse events following PCI.⁴ Moreover, most major complications occur within the first 6 hours after the procedure.⁵ These factors suggest that same-day discharge (SDD) following uncomplicated PCI is feasible when patients are carefully selected and adequately monitored for a sufficient period to ensure safety.⁴ Furthermore, implementing SDD may optimise resource usage, enhance patient comfort and potentially reduce complications associated with prolonged hospitalisation.^4,6 However, despite these potential benefits, SDD after uncomplicated PCI has not been widely adopted, and exhibits considerable variation across different countries.⁷ This variability is likely due to concerns regarding the safety of SDD, given the limited availability of high-quality evidence supporting its use.

The present study aimed to assess the safety of same-day discharge compared with overnight stay in patients undergoing elective PCI through a systematic review and meta-analysis of randomised controlled trials (RCTs).

Methods

This systematic review and meta-analysis was conducted in accordance with the PRISMA guidelines (Supplementary Table 1).⁸ The study was registered with the International Prospective Register for Systematic Reviews (PROSPERO - CRD420251001051).

Search Strategy

We conducted a systematic search of the PubMed, Embase, Cochrane Library and ClinicalTrials.gov databases. No restrictions were applied regarding the start date or language, and search results were updated until 22 February 2025. The search strategy was based on the following keywords: ‘PCI’, ‘PTCA’, ‘angioplasty’, ‘percutaneous coronary intervention’, ‘revascularisation’, ‘same-day’, ‘early’, ‘discharge’, ‘day-case’, ‘day-care’, ‘outpatient’, ‘short stay’ and ‘ambulatory’. The search terms and detailed strategies for each database are provided in Supplementary Table 2. Additionally, references from selected studies and relevant reviews were screened to identify any additional relevant literature.

Study Selection

Studies identified through the systematic search were initially screened by one investigator (DVN) based on titles and abstracts after the removal of duplicates. The full-text review of the remaining studies was then independently conducted by two investigators (DVN, HTTN) to assess eligibility according to the predefined inclusion criteria. Any discrepancies were resolved through discussion with a third investigator (TNL).

To ensure the highest-quality evidence for this study, the inclusion criteria were: RCTs comparing SDD with ONS following PCI; and studies reporting data on outcomes of interest within 30 days post-PCI. The exclusion criteria were: studies that reported pooled outcome data for SDD and ONS without the ability to separate the two groups; multiple reports originating from the same trial data set.

In cases where there was a significant risk of cohort overlap, only the study with the larger cohort was included, unless different studies reported distinct outcomes, in which case both were considered; and case reports or case series with fewer than five patients.

Data Extraction

Two investigators (DVN, HTTN) independently extracted data from the included studies, including study characteristics, baseline characteristics and clinical outcomes. Any discrepancies were resolved through consensus, and a third investigator (NHN) reviewed the extracted data for accuracy. Additionally, corresponding authors were contacted via email to obtain any missing or supplementary information as needed.

Risk of Bias and Study Quality Assessment

Two investigators (DVN, HTTN) independently evaluated the quality of the included studies and assessed the risk of bias using the Cochrane Risk of Bias 2 tool.⁹ Any disagreements were resolved through consensus. The Cochrane Risk of Bias 2 tool consists of five domains, and the overall risk of bias for each study was classified as low risk, some concerns or high risk of bias.⁹

Outcomes of Interest

The primary outcomes of interest assessed within 30 days post-PCI included: major adverse cardiovascular events (MACE): a composite of death, MI, coronary revascularisation and stroke; major bleeding events; vascular complications related to the access site, including arteriovenous fistula, pseudoaneurysm, radial artery occlusion, major access site bleeding, localised infection requiring antibiotics, moderate-to-severe haematoma (>5 cm in size) or any complication requiring surgical intervention; hospital readmission; and unplanned hospital visits. Additionally, stroke was assessed as an individual secondary outcome, whereas bleeding (encompassing both major and clinically relevant minor bleeding events) was evaluated as a composite secondary outcome.

Statistical Analysis

Variations among the included studies in terms of eligibility criteria, pre-, intra- and post-procedural protocols, study period, and geographic populations suggest that a fixed effects model would be insufficient to account for the heterogeneity of effect sizes.¹⁰ Therefore, we employed a random effects model using the DerSimonian–Laird method, with continuity corrections applied in cases of rare events. The difference in outcomes between the two strategies was estimated and reported as a RR with a 95% CI. Subgroup analyses were performed to evaluate the impact on clinical outcomes of the vascular access site (transfemoral versus transradial approach), the timing of discharge (<6 versus ≥6 hours post-PCI) and clinical presentation (including only chronic coronary syndrome [CCS] patients versus including both CCS and low-risk acute coronary syndrome [ACS] patients).

We assessed heterogeneity using Higgins’ I² statistic. An I² value <25% was considered low heterogeneity, 25–50% was classified as moderate heterogeneity and >50% indicated substantial heterogeneity.¹¹ A funnel plot and Egger’s test were used to assess publication bias. Sensitivity analysis was performed using the leave-one-out method to evaluate the impact of each individual trial on the overall study results.

All statistical analyses and plot generation were performed using Stata (version 18.0; StataCorp). Statistical significance was defined as p<0.05 (two-sided).

Results

Systematic Review of Published Studies

A total of 14 RCTs, comprising 3,215 patients (1,608 in the SDD group and 1,607 in the ONS group), were identified through our systematic search and included in this meta-analysis (Figure 1). Of these, nine trials exclusively included patients with CCS, one trial enrolled only low-risk ACS patients, and three trials included both CCS and a small proportion of low-to-moderate-risk ACS patients (Table 1).^12–24 Additionally, only one study included both CCS and a substantial proportion of ACS patients across low-, moderate- and high-risk categories, with ACS patients comprising 66% of the study cohort.²⁵ The quality assessment results are provided in Supplementary Figure 1, with none of the 14 RCTs classified as having a high risk of bias.

The mean age of the study populations ranged from 54 to 65 years, and the proportion of male patients varied between 63.3 and 94.9% (Table 1). The detailed medical history and comorbidities of participants in each study are presented in Supplementary Table 3. The prevalence of diabetes ranged from 15.1 to 49.3%, hypertension from 53.6 to 90.9% and dyslipidaemia from 25.4 to 90.9%. Regarding coronary history, the proportion of patients with a previous MI ranged from 25 to 60%, prior PCI from 7 to 52.2% and prior coronary artery bypass grafting from 2 to 13.3%.

Although patient selection criteria varied across studies (Supplementary Table 4), most trials shared common eligibility criteria, including elective PCI, absence of high-risk features pre-PCI (e.g. advanced age, moderate-to-severe renal impairment, severe heart failure, coagulopathy or complex comorbidities), successful PCI with optimal angiographic results post-stent implantation, no procedural complications and clinically stable status post-PCI. Additionally, eligible patients were required to have a preference for early discharge, with adequate familial, geographic and healthcare conditions to allow timely access to an emergency centre within 40–60 minutes in case of complications.

The detailed characteristics of percutaneous coronary procedures among participants in the included studies are presented in Supplementary Table 5. Among the included studies, eight trials reported the use of the transfemoral approach, with five of them using vascular closure devices. Sheath sizes ranged from 5 to 8 Fr, depending on the study and vascular access site. The proportion of patients undergoing multivessel PCI varied from 12.2 to 33% across studies. The post-PCI observation period before discharge in the SDD group ranged from 4 to 12 hours, with the majority of trials (6 studies) suggesting that a 6-hour monitoring period was optimal to ensure patient safety following discharge.

The definitions used for ischaemic and bleeding endpoints (MI, major bleeding, and minor bleeding) are provided in Supplementary Table 6. In general, criteria for myocardial infarction across the included trials required a significant rise in cardiac biomarkers (creatine kinase and troponin), whereas definitions of major bleeding events predominantly involved a substantial drop in haemoglobin levels and the necessity for blood product transfusion. However, considerable variability existed among studies with respect to the biomarker and haemoglobin cut-off values.

Outcomes of Interest

Overall, our meta-analysis indicates that within 30 days post-PCI, there was no statistically significant difference in the risk of MACE between SDD and ONS strategies (pooled RR 0.76; 95% CI [0.46–1.27]; p=0.30; Figure 2). Likewise, major bleeding events (pooled RR 1.29; 95% CI [0.50–3.37]; p=0.60; Figure 3) and vascular complications related to the access site (pooled RR 1.06; 95% CI [0.78–1.45]; p=0.70; Figure 2) did not differ between the two approaches. Furthermore, the composite bleeding endpoint, which included both major and clinically relevant minor events, also revealed comparable risks between SDD and ONS (pooled RR 1.12; 95% CI [0.82–1.52]; p=0.47; Supplementary Figure 2).

Rehospitalisation (pooled RR 1.15; 95% CI [0.79–1.68]) and unplanned hospital visit rates (pooled RR 1.02; 95 % CI [0.73–1.42]) were also similar between groups (Figure 4). Additionally, when evaluated as an individual secondary outcome, the incidence of stroke within 30 days post-PCI was equivalent between SDD and ONS (pooled RR 0.99; 95 % CI [0.17–5.64]; p=0.99; Supplementary Figure 3). Moreover, the very low heterogeneity among studies (I²=0 for all outcomes of interest) indicates a high level of consistency in the existing evidence.

Sensitivity analysis demonstrated that the comparison of adverse event risks between the SDD and ONS strategies remained robust, as the pooled RR showed minimal variation across all outcomes of interest in the leave-one-out analysis (Supplementary Figure 4).

Subgroup Analysis Based on Vascular Access Strategy

The results of the subgroup analyses, presented in Supplementary Figure 5, revealed no significant association between the vascular access site (transradial versus transfemoral) and the effect sizes of the outcomes of interest (p for interaction [p_in]=0.82 for MACE; p_in=0.45 for major bleeding; p_in=0.58 for composite bleeding endpoint [both major and minor bleeding]; p_in=0.86 for vascular complications; p_in=0.22 for rehospitalisation; p_in=0.48 for unplanned hospital visits; p_in=0.97 for stroke). Similar findings were observed in the stratified analyses based on discharge timing (discharged at <6 versus ≥6 hours post-PCI; p_in=0.70 for MACE; p_in=0.35 for major bleeding; p_in=0.82 for composite bleeding endpoint; p_in=0.82 for vascular complications; p_in=0.81 for rehospitalisation; p_in=0.26 for unplanned hospital visits; p_in=0.97 for stroke; Supplementary Figure 6) and clinical presentation (including only CCS patients versus including both CCS and ACS patients; p_in=0.62 for MACE; p_in=0.27 for major bleeding; p_in=0.50 for composite bleeding endpoint; p_in=0.64 for vascular complications; p_in=0.30 for rehospitalisation; p_in=0.72 for unplanned hospital visits; p_in=0.98 for stroke; Supplementary Figure 7).

Publication Bias

The assessment of publication bias for the outcomes of interest demonstrated symmetric funnel plots, with no clear evidence of small-study effects (Supplementary Figure 8). This finding was further supported by non-significant Egger’s test results (p>0.05).

Discussion

To the best of our knowledge, this study represents the most up-to-date and largest systematic review and meta-analysis of RCTs comparing SDD and ONS, including 14 trials with a total of 3,215 patients. The pooled analysis demonstrated that SDD is as safe as ONS, as evidenced by the comparable risk of adverse events, including MACE, major bleeding, vascular complications, rehospitalisation and unplanned hospital visits within 30 days post-PCI. Furthermore, the evidence supporting SDD safety was consistent across studies, regardless of whether the transradial or transfemoral approach was used.

The feasibility of SDD after PCI has increased over the years, largely due to significant improvements in short-term cardiovascular outcomes following PCI.⁴ This progress is primarily attributed to advancements in vascular access techniques, antithrombotic/antiplatelet therapies and drug-eluting stent (DES) technology.^4,26 The increasing implementation of guideline-directed antiplatelet therapy, with clopidogrel for elective, CCS and non-complex PCI patients, and prasugrel or ticagrelor for ACS and complex PCI patients, has led to a better balance between the risk of ischaemic events and bleeding complications in the management of post-PCI patients, thereby enhancing patient safety.^4,27 Furthermore, DES technology has evolved significantly in recent years, with improvements in stent expansion and flexibility, thinner and more biocompatible polymers, and newer antiproliferative drugs.^28,29 These second-generation DES advancements have contributed to a lower risk of late lumen loss and stent thrombosis, leading to reduced restenosis rates and superior long-term outcomes compared with bare-metal stents and first-generation DES.^28,29 Given these advancements, the periprocedural complication rate following PCI has markedly improved, suggesting that routine overnight observation after an uncomplicated PCI in stable CAD patients may not be necessary.^2,4

In this study, the pooled incidence of MACE within 30 days post-PCI across the 14 included trials was estimated at 1.3% in the SDD group and 1.8% in the ONS group, with no significant difference between the two strategies (pooled RR 0.76; 95% CI [0.46–1.27]). Additionally, improvements in periprocedural anticoagulation strategies, and enhanced early detection and prevention of bleeding risks have contributed to the very low incidence of major bleeding events (<1%) in the RCTs included in this review.² Importantly, a non-significant difference in major bleeding between the SDD and ONS groups (pooled RR 1.29; 95% CI [0.50–3.37]) suggests that SDD does not increase short-term bleeding risks compared with ONS after PCI.

Beyond concerns regarding thrombotic events, vascular complications at the access site remain one of the primary reasons for prolonged post-PCI monitoring. However, reports indicate that the incidence of access-site complications has steadily declined, primarily due to the adoption of the transradial approach; advancements in femoral access techniques and sheath removal protocols; the use of smaller sheath sizes; and the implementation of vascular closure devices for femoral access haemostasis, as well as a combination of these measures.³⁰

In our study, the majority of trials (eight studies) used the transfemoral approach, and among these, five trials used vascular closure devices. Meanwhile, five other trials exclusively included patients who underwent PCI via transradial approach. The transradial approach has been well-established as being associated with significantly lower vascular complications and bleeding risk compared with the transfemoral approach, while also facilitating faster post-PCI recovery.^26,31

However, the findings from previous studies indicate that the choice of vascular access site (radial versus femoral) does not significantly affect the safety or success of the SDD strategy.^2,30 This is likely due to improvements in the safety of the transfemoral approach in recent years.³² The incidence of access-related complications has declined to approximately 2% in recent studies.^32–34 This improvement in transfemoral approach safety is largely attributed to refinements in arterial access techniques (including routine fluoroscopic guidance, micropuncture techniques and ultrasound guidance), advancements in pharmacological therapy, reductions in sheath size and modifications in sheath removal techniques.^32,34–36 For instance, reducing sheath size has been shown to decrease access-site bleeding and vascular complications by nearly 50%.³⁷

In contrast, although vascular closure devices provide haemostasis comparable to manual compression in terms of vascular complications and bleeding risk, they facilitate earlier ambulation post-procedure and enhance the feasibility of implementing the SDD strategy.³⁸ In the current meta-analysis, there was no significant difference in the risk of major bleeding or vascular complications related to the access site between the transradial approach and transfemoral approach subgroups. However, these findings may be confounded by limited sample sizes and the inclusion of predominantly low-risk patients.

Given the well-documented association between femoral access and an increased risk of vascular complications (particularly among patients requiring crossover from radial to femoral access, who have been shown to experience a higher incidence of access site-related major bleeding compared with those with successful radial access), recommendations for early discharge in this subgroup should be made with caution.³⁹ Clinical decisions should involve careful patient selection and individualised assessment by treating physicians.

Risk stratification for prognostication following PCI is critical in determining the necessary length of hospital stay. Selecting patients without high-risk clinical features enhances both the safety and feasibility of the SDD strategy. Likely for this reason, although patient selection criteria vary across studies, most trials focused on low-risk populations with key eligibility criteria, including CCS or low-risk ACS; elective PCI procedures; absence of high-risk clinical factors prior to PCI (e.g. advanced age, moderate-to-severe renal dysfunction, severe heart failure, coagulopathy, complex comorbidities); successful PCI with optimal post-stenting angiographic results; no periprocedural complications; clinically stable condition post-PCI; patient preference for early discharge; and adequate familial, geographic and healthcare support to ensure timely access to emergency care (<40–60 minutes post-discharge) in case of complications.

In the present study, stratified meta-analysis revealed comparable outcomes between trials that included only CCS patients and those that included both CCS and low-risk ACS patients, suggesting that the SDD strategy is appropriate for both these populations. Nevertheless, as most included studies excluded patients with complex PCI – such as left main disease, proximal left anterior descending coronary artery lesions, saphenous vein graft lesions, type C lesions and chronic total occlusions – the study population predominantly comprised individuals undergoing low-complexity PCI. Consequently, the generalisability of our findings to high-complexity PCI populations is limited.

The primary purpose of the post-PCI monitoring period is to detect and manage potential latent complications, particularly bleeding, access-site complications, acute in-stent thrombosis and recurrent ischaemia.² Notably, findings from previous studies demonstrated that all recorded complications occurred either within the first 6 hours post-PCI or beyond 24 hours.^40,41 These findings support the hypothesis that most adverse events occur within the first 6 hours or after 24 hours, suggesting that routine overnight observation may not be necessary. Notably, our stratified analysis comparing trials that discharged patients within <6 hours post-PCI with those that discharged patients at ≥6 hours revealed no significant differences in outcomes between the two groups. Based on these findings, a standard observation period of 6 hours may be appropriate for general clinical practice. However, this duration should be individualised and potentially shortened depending on the physician’s assessment, considering the procedural success and the patient’s clinical condition.

Despite the evidence supporting the safety of the SDD strategy observed across the trials included in our review, its adoption remains low in most clinical centres.⁴² To facilitate broader implementation of SDD, in addition to applying the strategy based on the evidence provided by this systematic review, it is essential to address other potential barriers contributing to its limited uptake. These include the lack of financial incentives for healthcare providers, the relatively low adoption rate of transradial access, and inconsistencies in funding and resource allocation within universal healthcare systems.^42–44

Study Limitations

First, although we made every effort to include all relevant and up-to-date trials, the modest sample size and low event rates may limit the statistical power of the evidence. Second, most included trials enrolled carefully selected patients undergoing elective PCI, primarily those with CCS and low-risk ACS, which may limit the generalisability of our findings to intermediate- and high-risk ACS populations. Third, acute kidney injury is a short-term complication following PCI that warrants consideration; however, it was not assessed or reported in the included trials. Nevertheless, since acute kidney injury typically manifests several days after PCI, the decision to discharge patients on the same day or extend observation overnight is unlikely to have a substantial impact on its occurrence.² Fourth, the lack of individual participant data related to procedural characteristics (e.g. procedure type, number of stents deployed, lesion type) limits our ability to perform further analyses to assess the impact of procedural complexity on the safety of SDD. Fifth, the inconsistency in endpoint definitions across trials may introduce significant heterogeneity into this meta-analysis and, consequently, warrants cautious interpretation of the results. Finally, as the study population predominantly consisted of male patients, the generalisability of the findings to female patients is limited. This is particularly noteworthy given that a recent study involving 16,821 patients demonstrated a more than twofold higher incidence of access-site bleeding in women undergoing PCI.⁴⁵ Consequently, ensuring effective haemostasis and considering a longer post-procedural observation period may be necessary to implement a safe and successful SDD strategy for female patients.

Conclusion

An up-to-date meta-analysis of 14 RCTs, primarily including patients with CCS and low-risk acute ACS who underwent PCI, demonstrated that the SDD strategy is as safe as routine overnight monitoring regarding MACE, major bleeding, vascular complications, rehospitalisation and unplanned hospital visits. Further large-scale RCTs involving high-risk ACS patients are warranted to further clarify the safety and feasibility of the SDD strategy following PCI in this population.

Click here to view Supplementary Material.