Radial access for percutaneous coronary intervention (PCI) has been shown to reduce mortality and vascular complications compared to femoral access in patients with ST-segment elevation myocardial infarction. However, efficacy and safety of radial access PCI in non–ST-segment elevation acute coronary syndrome (NSTE ACS) is not well understood. A systematic search of electronic databases was performed through July 2015 to search and identify relevant studies. We evaluated the following short-term outcomes: all-cause mortality, major bleeding, access site bleeding, and need for blood transfusions. In addition, we evaluated 1-year mortality. Studies were pooled using random effects model. Nine studies including a total of 220,126 patients (radial approach: 94,663 patients [43%], femoral approach: 125,463 patients [57%]) were included in the analysis. On pooled analysis, no significant difference in incidence of short-term all-cause mortality was found between radial and femoral access (odds ratio [OR] 0.78, 95% CI 0.57 to 1.07, p = 0.12). Radial access was associated with significant reduction in major bleeding (OR 0.52, 95% CI 0.36 to 0.73, p = 0.0002), access-site bleeding (OR 0.41, 95% CI 0.22 to 0.78, p = 0.007), and need for blood transfusions (OR 0.61, 95% CI 0.41 to 0.91, p = 0.02). Furthermore, the 1-year mortality was significantly lower in radial approach (OR 0.72, 95% CI 0.55 to 0.95, p = 0.02). In conclusion, in patients with non–ST-segment elevation acute coronary syndrome undergoing PCI, radial access is associated with decreased bleeding and access-site complications.
Non–ST-segment elevation acute coronary syndrome (NSTE ACS) represents the most common presentation for ACS. Invasive strategy using percutaneous coronary intervention (PCI) is associated with improved clinical outcomes in patients with NSTE ACS. Patients undergoing invasive strategy receive multiple antithrombotic therapies, which increases risk for bleeding complications. In patients undergoing PCI, radial access has been shown to reduce access site–related bleeding, vascular complications, and mortality compared with femoral access. However, most of the evidence favoring radial access comes from randomized control trials (RCTs) of patients with ST-segment elevation myocardial infarction (STEMI) or ACSs in general. There is no RCT that has studied the effect of radial versus femoral access specifically in patients with NSTE ACS. Currently, the only randomized data available are from subgroup analysis of Radial Versus Femoral Access for Coronary Intervention (RIVAL) trial and the recently published Minimizing Adverse Haemorrhagic Events by Transradial Access Site and Systemic Implementation of angiox (MATRIX) trial. However, both the trials were underpowered to study differences in outcomes in NSTE ACS cohort. In view of conflicting results and lack of adequate data from RCTs, we undertook a systemic review and meta-analysis of all available studies comparing radial with femoral access in patients with NSTE ACS to study the effect of radial access on clinical outcomes.
Methods
We performed a systematic search, without language restriction, using PubMed, EMBASE, Scopus, Web of Science, ClinicalTrials.gov , Cochrane library, and Google Scholar databases, from inception to July 2015, to identify all the studies comparing radial versus femoral access for PCI in patients with NSTE ACS. We also searched conference proceedings of the following societies: Transcatheter Cardiovascular Therapeutics, Euro-PCR, Society of Cardiovascular Angiography and Intervention, American College of Cardiology, American Heart Association, and European Society of Cardiology, presented in the last 10 years. Furthermore, we performed manual searches through the reference lists of studies and pertinent reviews on this topic. The search keywords included the following MeSH terms: “femoral” OR “radial” OR “transradial” OR “access” OR “transfemoral” OR “PCI” AND “ACS” OR “unstable angina” OR “non-ST-segment elevation myocardial infarction” OR “NSTEMI” OR “NSTE ACS.” When there were multiple reports from the same database/patient cohort, we used the most complete and relevant study. Two researchers (CB and SP) independently and in duplicate performed the literature search.
Studies were included if they met the following inclusion criteria: (1) RCT or observational studies comparing radial versus femoral access for PCI in patients with NSTE ACS; and (2) studies providing short-term or long-term outcomes particularly mortality and/or bleeding outcomes. Studies were excluded if they evaluated radial and femoral access exclusively in patients with STEMI or in ACS but not reporting any outcomes data for NSTE ACS cohort.
Two physician-reviewers (CB and SP) independently reviewed the originally identified titles and abstracts and selected studies for pooled analysis on the basis of the inclusion and exclusion criteria. We extracted the following data from individual studies: first investigator, year of publication, study enrollment period, design of the study, sample size for radial and femoral access group, patient characteristics, length of follow-up, mortality, cardiovascular, and bleeding outcomes data. Any divergence was resolved with consensus. For RCT, quality and assessment of trial bias risk was evaluated per Cochrane collaboration criteria, specifically emphasizing sequence generation, allocation concealment, blinding, outcomes assessment, and selective reporting. The quality of observational studies was assessed using the Newcastle-Ottawa Scale.
The following short-term outcomes were evaluated: all-cause mortality, bleeding, access-site bleeding, and blood transfusion. Short-term mortality included in-hospital or 30-day all-cause mortality. The definition of major bleeding followed the bleeding definition used in each study. When a study reported multiple bleeding outcomes using various criteria, Thrombolysis In Myocardial Infarction non–coronary artery bypass grafting major bleeding was used for the analysis. Major adverse cardiac event was not evaluated because of variable definitions used in different studies. Of long-term outcomes, 1-year mortality was evaluated because it was the only outcome that was consistently reported in included studies.
The statistical analysis was done in line with recommendations from the Meta-analysis of Observational Studies in Epidemiology and the Preferred Reporting Items for Systematic reviews and Meta-analyses guidelines. Because of known methodological and statistical heterogeneity, we used the random-effects (DerSimonian and Laird) model to pool effect estimates. Heterogeneity was assessed using Higgins and Thompson’s I 2 statistic. I 2 is the proportion of total variation observed between the trials attributable to differences between trials rather than sampling error (chance) with I 2 values of <25%, 25% to 75%, and >75% correspond to low, moderate, and high levels of heterogeneity, respectively. We also performed meta-influence analysis, whereby one study was removed at a time from the pool analysis to assess whether any study has a significant influence on the overall pooled analysis. Publication bias was estimated visually by funnel plots and validated using the Egger’s weighted regression test. In case of any publication bias, Duval and Tweedie’s trim-and-fill method was used to adjust for publication bias. A 2-tailed p <0.05 was considered statistically significant for all the analyses. Statistical analysis was performed using RevMan, version 5.02, (Copenhagen: The Nordic Cochrane Center, The Cochrane Collaboration, 2014) and Stata 11 (Stata Corp., College Station, Texas).
Results
Our initial search, after removing duplicates, yielded 967 citations of which 9 studies were included for the final analysis on the basis of our inclusion and exclusion criteria ( Figure 1 ). Characteristics of included studies are summarized in Table 1 . Selected studies include 2 RCTs (RIVAL and MATRIX trials), post hoc analyses from 3 RCTs and 4 retrospective cohort studies. All studies were multicentered studies. The included studies contained a total of 220,126 patients, with 94,663 patients (43%) in radial access group and 125,463 patients (57%) in femoral access group. All the studies reported short-term (in-hospital or 30-day) outcomes with 5 studies also reporting long-term (1-year) mortality. The characteristics of patients in the included studies are presented in Table 2 . Most studies involved heparin only or heparin ± glycoprotein IIb to IIIa inhibitors, whereas only 3 studies reported the use of bivalirudin. All the studies were of good quality to be included in the quantitative analysis ( Supplementary Table 1 ).
| First Author | Enrollment | Design | Radial | Femoral | Outcomes | Bleeding definitions used |
|---|---|---|---|---|---|---|
| Cantor, 2006 | 2001-2003 | Post-hoc analysis of SYNERGY | 413 | 8,922 | In-hospital | Non-CABG TIMI major bleeding |
| Hamon, 2009 | 2003-2006 | Post-hoc analysis of ACUITY | 798 | 11,989 | 30-day/1-year | Non-CABG TIMI major bleeding |
| Iqbal, 2014 | 2005-2011 | Retrospective cohort study | 2,275 | 7,820 | 30-day/1-year | Access-site bleeding, gastrointestinal bleeding, cardiac tamponade, intracranial hemorrhage, blood transfusion |
| Klutstein, 2013 | 2004-2008 | Post-hoc analysis of EARLY-ACS | 1,230 | 7,896 | 30-day/1-year | TIMI major bleeding |
| Mehta, 2012 | 2006-2011 | RIVAL trial | 2,552 | 2,511 | 30-day | Non-CABG major bleeding (study definition) |
| Park, 2013 | Jan-Dec 2009 | Retrospective cohort study | 402 | 593 | In-hospital/ 1-year | CRUSADE major bleeding |
| Ratib, 2015 | 2007-2012 | Retrospective cohort study | 84,490 | 82,671 | 30-day | Any gastrointestinal bleed, intracerebral bleed, retroperitoneal bleed, or blood transfusion |
| Sciahbasi, 2009 | 2003-2006 | Retrospective cohort study | 307 | 863 | In-hospital/ 1-year | TIMI major and minor bleeding |
| Valgimigli, 2015 | 2011-2014 | MATRIX trial | 2196 | 2198 | 30-day | BARC type 3 or 5 bleeding |
| First Author | Age (years) | Male (%) | Diabetes (%) | CKI (%) | UFH (%) | LMWH (%) | GpI (%) | Bival (%) | Stent (%) | Sheath (≥7F) |
|---|---|---|---|---|---|---|---|---|---|---|
| Cantor | 66/67 | 72/66 ∗ | 27/29 | NR | 50/27 ∗ | 47/50 | 59/59 | NR | 78/71 | 4/20 ∗ |
| Hamon | 61/63 ∗ | 76/69 ∗ | 21/28 ∗ | 19/19 | 35/33 † | NR | 32/33 | 65/67 | NR | NR |
| Iqbal | 65/66 ∗ | 78/63 ∗ | 23/23 | 2/3 | NR | NR | 29/30 | 0.5/0.3 | NR | NR |
| Klutstein | 65/68 ∗ | 74/68 ∗ | 28/31 ∗ | 15/18 ∗ | 32/48 ∗ | 74/64 ∗ | NR | NR | 61/60 | NR |
| Mehta | 63/63 | 72/71 | 24/22 | NR | 43/43 | 55/55 | 22/21 | 2/3 | 95/95 | 1/7 ∗ |
| Park | 72/71 | 34/37 | 50/53 | 74/80 ∗ | 42/69 ∗ | 18/5 ∗ | 2/6 ∗ | NR | NR | 12/91 ∗ |
| Ratib | NR | NR | NR | NR | NR | NR | NR | NR | NR | NR |
| Sciahbasi | 65/68 ∗ | 73/66 ∗ | 27/31 | NR | 34/30 | 67/71 | 52/34 ∗ | NR | 67/61 | NR |
| Valgimigli | NR | NR | NR | NR | NR | NR | NR | NR | NR | NR |
On pooled analysis, no significant difference in incidence of short-term mortality was found between radial and femoral access (odds ratio [OR] 0.78, 95% CI 0.57 to 1.07, p = 0.12; Figure 2 ). Radial access was associated with 48% decrease in major bleeding (OR 0.52, 95% CI 0.36 to 0.73, p = 0.0002), 69% decrease in access-site bleeding (OR 0.41, 95% CI 0.22 to 0.78, p = 0.007), and 39% decrease in the need for blood transfusions (OR 0.61, 95% CI 0.41 to 0.91, p = 0.02) compared with femoral access in patients with NSTE ACS ( Table 3 ). When studies using only standardized bleeding definition were analyzed, radial access showed significant reduction in major bleeding (OR 0.67, 95% CI 0.53 to 0.84, p = 0.0005). Furthermore, the 1-year mortality was significantly lower in radial group than in femoral group (OR 0.72, 95% CI 0.55 to 0.95, p = 0.02; Figure 3 ). A moderate-to-high heterogeneity was observed in all the analysis. Visual inspection of funnel plots and Egger’s weighted regression test showed no obvious evidence of publication bias for short-term outcomes ( Supplementary Figure 1 ). This was confirmed using Duval and Tweedie’s trim-and-fill method, in which no studies were imputed.
