Transradial versus transfemoral approach for diagnostic coronary angiography and percutaneous coronary intervention in people with coronary artery disease

Ahmed A Kolkailah; Rabah S Alreshq; Ahmed M Muhammed; Mohamed E Zahran; Marwah Anas El‐Wegoud; Ashraf F Nabhan

doi:10.1002/14651858.CD012318.pub2

. 2018 Apr 18;2018(4):CD012318. doi: 10.1002/14651858.CD012318.pub2

Transradial versus transfemoral approach for diagnostic coronary angiography and percutaneous coronary intervention in people with coronary artery disease

Ahmed A Kolkailah ^1,^✉, Rabah S Alreshq ², Ahmed M Muhammed ³, Mohamed E Zahran ³, Marwah Anas El‐Wegoud ⁴, Ashraf F Nabhan ⁵

Editor: Cochrane Heart Group

PMCID: PMC6494633 PMID: 29665617

Abstract

Background

Cardiovascular disease (CVD) is the major cause of mortality worldwide. Coronary artery disease (CAD) contributes to half of mortalities caused by CVD. The mainstay of management of CAD is medical therapy and revascularisation. Revascularisation can be achieved via coronary artery bypass grafting (CABG) or percutaneous coronary intervention (PCI). Peripheral arteries, such as the femoral or radial artery, provide the access to the coronary arteries to perform diagnostic or therapeutic (or both) procedures.

Objectives

To assess the benefits and harms of the transradial compared to the transfemoral approach in people with CAD undergoing diagnostic coronary angiography (CA) or PCI (or both).

Search methods

We searched the following databases for randomised controlled trials on 10 October 2017: Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, Embase, and Web of Science Core Collection. We also searched ClinicalTrials.gov and the World Health Organization International Clinical Trials Registry Platform in August 2017. There were no language restrictions. Reference lists were also checked and we contacted authors of included studies for further information.

Selection criteria

We included randomised controlled trials that compared transradial and transfemoral approaches in adults (18 years of age or older) undergoing diagnostic CA or PCI (or both) for CAD.

Data collection and analysis

We used the standard methodological procedures expected by Cochrane. At least two authors independently screened trials, extracted data, and assessed the risk of bias in the included studies. We contacted trial authors for missing information. We used risk ratio (RR) for dichotomous outcomes and mean difference (MD) or standardised mean difference (SMD) for continuous data, with their 95% confidence intervals (CIs). All analyses were checked by another author.

Main results

We identified 31 studies (44 reports) including 27,071 participants and two ongoing studies. The risk of bias in the studies was low or unclear for several domains. Compared to the transfemoral approach, the transradial approach reduced short‐term net adverse clinical events (NACE) (i.e. assessed during hospitalisation and up to 30 days of follow‐up) (RR 0.76, 95% CI 0.61 to 0.94; 17,133 participants; 4 studies; moderate quality evidence), cardiac death (RR 0.69, 95% CI 0.54 to 0.88; 11,170 participants; 11 studies; moderate quality evidence). However, short‐term myocardial infarction was similar between both groups (RR 0.91, 95% CI 0.81 to 1.02; 19,430 participants; 11 studies; high quality evidence). The transradial approach had a lower procedural success rate (RR 0.97, 95% CI 0.96 to 0.98; 25,920 participants; 28 studies; moderate quality evidence), but was associated with a lower risk of all‐cause mortality (RR 0.77, 95% CI 0.62 to 0.95; 18,955 participants; 10 studies; high quality evidence), bleeding (RR 0.54, 95% CI 0.40 to 0.74; 23,043 participants; 20 studies; low quality evidence), and access site complications (RR 0.36, 95% CI 0.22 to 0.59; 16,112 participants; 24 studies; low quality evidence).

Authors' conclusions

Transradial approach for diagnostic CA or PCI (or both) in CAD may reduce short‐term NACE, cardiac death, all‐cause mortality, bleeding, and access site complications. There is insufficient evidence regarding the long‐term clinical outcomes (i.e. beyond 30 days of follow‐up).

Plain language summary

Radial artery versus femoral artery approach for performing coronary catheter procedures in people with coronary artery disease

Review question

Should physicians introduce the catheter (a long, thin tube) through the femoral artery (transfemoral access via the groin) or the radial artery (transradial access via the wrist) to reach the coronary arteries (blood vessels supplying the heart) for the diagnosis or treatment of coronary artery disease?

Background

Coronary artery disease contributes to half of deaths caused by cardiovascular (heart and blood vessels) disease. Restoration of adequate blood flow through the coronary arteries can be achieved by introducing a catheter through a peripheral artery. This allows the introduction of balloons through the aorta (major artery of the heart) to dilate coronary artery narrowing or place arterial scaffolds (tubes called stents) to keep the coronary arteries open. Two main peripheral arteries can provide access; traditionally, the femoral (groin) artery, and more recently, the radial artery (one of two major arteries in the forearm). While gaining popularity, the transradial approach can be more challenging than the transfemoral approach, which may translate to longer procedural durations and technical failures. In addition, this raises concerns regarding radiation exposure to patients and physicians being higher with the transradial approach. We sought to compare the advantages and disadvantages of both approaches to help inform healthcare decisions.

Study characteristics

Our search yielded 31 eligible studies comparing the transradial approach to the transfemoral approach in people undergoing diagnostic or therapeutic (or both) coronary catheterisation procedures in different settings, whether urgent (during heart attacks (myocardial infarctions)) or elective (planned procedure). The trials were carried out in many countries and regions, including Canada, China, Europe, Japan, and USA. We also identified two ongoing studies. The evidence was current to October 2017.

Key results

Transradial access was associated with a reduction in the composite outcome (comprising two or more combined outcomes) of net adverse clinical events (NACE), including death from cardiac causes, myocardial infarction (injury of the heart muscle), stroke (insult to the brain), need to reintervene on the same site of coronary artery stenosis (narrowing), and bleeding during the first 30 days following intervention. When assessing individual outcomes, the risk of myocardial infarction and stroke was similar between groups. Transradial access reduced death from cardiac causes, death from all causes during the first 30 days following intervention, bleeding, and local complications at the access site. The transradial approach shortened the length of stay in hospital, but was associated with a higher radiation exposure and more technical failures requiring an alternate vascular access route.

Quality of the evidence

We rated the quality of the evidence for short‐term myocardial infarction and all‐cause death as high. We rated short‐term NACE, cardiac death, and success of the procedure as moderate quality evidence. Evidence for bleeding and access site complications was low quality.

Summary of findings

Summary of findings for the main comparison. Transradial compared to transfemoral approach for diagnostic coronary angiography and percutaneous coronary intervention in people with coronary artery disease.

Transradial compared to transfemoral approach for diagnostic coronary angiography and percutaneous coronary intervention in people with coronary artery disease
Patient or population: people with coronary artery disease undergoing diagnostic coronary angiography and percutaneous coronary intervention  Setting: inpatient  Intervention: transradial  Comparison: transfemoral
Outcomes¹	Anticipated absolute effects² (95% CI)		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Short‐term NACE	90 per 1000	68 per 1000  (55 to 85)	RR 0.76  (0.61 to 0.94)	17,133  (4 RCTs)	⊕⊕⊕⊝  Moderate³	‐
Short‐term cardiac death	26 per 1000	18 per 1000  (14 to 23)	RR 0.69  (0.54 to 0.88)	11,170  (11 RCTs)	⊕⊕⊕⊝  Moderate⁴	‐
Short‐term MI	55 per 1000	50 per 1000  (45 to 56)	RR 0.91  (0.81 to 1.02)	19,430  (11 studies)	⊕⊕⊕⊕  High	Although 15 studies reported this outcome, 4 had 0 events in either group, so the meta‐analysis was based on 11 studies.
Success of the procedure	979 per 1000	950 per 1000  (940 to 960)	RR 0.97  (0.96 to 0.98)	25,920  (28 RCTs)	⊕⊕⊕⊝  Moderate³	‐
Short‐term all‐cause mortality	18 per 1000	14 per 1000  (11 to 17)	RR 0.77  (0.62 to 0.95)	18,955  (10 RCTs)	⊕⊕⊕⊕  High	‐
Bleeding	98 per 1000	53 per 1000  (39 to 73)	RR 0.54  (0.40 to 0.74)	23,043  (20 RCTs)	⊕⊕⊝⊝  Low^3,4	Although 22 studies reported this outcome, 2 had 0 events in either group, so the meta‐analysis was based on 20 studies.
Access site complications	50 per 1000	18 per 1000  (11 to 30)	RR 0.36  (0.22 to 0.59)	16,112  (24 RCTs)	⊕⊕⊝⊝  Low^3,4	‐
CI: confidence interval; MI: myocardial infarction; NACE: net adverse clinical events; RCTs: randomised controlled trials; RR: risk ratio.
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Short‐term NACE	4	17133	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.61, 0.94]
2 Short‐term NACE (CA vs PCI)	3	8729	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.55, 0.98]
2.1 CA	1	2361	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.66, 1.44]
2.2 PCI	3	6368	Risk Ratio (M‐H, Random, 95% CI)	0.67 [0.48, 0.95]
3 Short‐term NACE (STEMI vs NSTE‐ACS)	4	17133	Risk Ratio (M‐H, Random, 95% CI)	0.77 [0.64, 0.93]
3.1 STEMI	4	7676	Risk Ratio (M‐H, Random, 95% CI)	0.67 [0.51, 0.87]
3.2 NSTE‐ACS	2	9457	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.71, 1.23]
4 Short‐term NACE (women vs men)	2	15515	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.76, 0.95]
4.1 Women	2	4093	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.62, 0.92]
4.2 Men	2	11422	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.79, 1.02]
5 Short‐term cardiac death	11	11170	Risk Ratio (M‐H, Fixed, 95% CI)	0.69 [0.54, 0.88]
6 Short‐term cardiac death (STEMI vs NSTE‐ACS)	1	8404	Risk Ratio (M‐H, Fixed, 95% CI)	0.76 [0.55, 1.04]
6.1 STEMI	1	4010	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.59, 1.29]
6.2 NSTE‐ACS	1	4394	Risk Ratio (M‐H, Fixed, 95% CI)	0.56 [0.32, 1.00]
7 Short‐term MI	11	19430	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.81, 1.02]
8 Short‐term MI (elective vs primary PCI)	9	18409	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.80, 1.04]
8.1 Elective	2	712	Risk Ratio (M‐H, Fixed, 95% CI)	1.2 [0.53, 2.74]
8.2 Primary PCI	7	17697	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.80, 1.03]
9 Short‐term MI (STEMI vs NSTE‐ACS)	1	8404	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.78, 1.05]
9.1 STEMI	1	4010	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.73, 1.43]
9.2 NSTE‐ACS	1	4394	Risk Ratio (M‐H, Fixed, 95% CI)	0.88 [0.75, 1.04]
10 Short‐term MI (sensitivity analysis)	6	18377	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.81, 1.05]
11 Long‐term MI	3	556	Risk Ratio (M‐H, Fixed, 95% CI)	1.77 [0.38, 8.20]
12 Success of the procedure	28	25920	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.96, 0.98]
13 Success of the procedure (CA vs PCI)	14	5787	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.94, 0.98]
13.1 CA	2	890	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.94, 1.01]
13.2 PCI	13	4897	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.94, 0.98]
14 Success of the procedure (elective vs primary PCI)	19	20052	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.95, 0.98]
14.1 Elective	4	987	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.92, 1.02]
14.2 Primary PCI	15	19065	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.95, 0.98]
15 Success of the procedure (STEMI vs NSTE‐ACS)	11	3248	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.95, 0.99]
15.1 STEMI	10	3008	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.95, 0.99]
15.2 NSTE‐ACS	1	240	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.93, 1.02]
16 Success of the procedure (women vs men)	2	1887	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.93, 0.97]
16.1 Women	1	1775	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.93, 0.97]
16.2 Men	1	112	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.82, 1.00]
17 Success of the procedure (sensitivity analysis)	11	21820	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.94, 0.96]
18 Short‐term all‐cause mortality	10	18955	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.62, 0.95]
19 Long‐term all‐cause mortality	3	1013	Risk Ratio (M‐H, Fixed, 95% CI)	0.62 [0.29, 1.32]
20 Bleeding	20	23043	Risk Ratio (M‐H, Random, 95% CI)	0.54 [0.40, 0.74]
21 Short‐term stroke	9	19017	Risk Ratio (M‐H, Fixed, 95% CI)	1.08 [0.74, 1.60]
22 Access site complications	24	16112	Risk Ratio (M‐H, Random, 95% CI)	0.36 [0.22, 0.59]
23 Total radiation dose	4	980	Std. Mean Difference (IV, Fixed, 95% CI)	0.19 [0.07, 0.32]
23.1 CA	2	321	Std. Mean Difference (IV, Fixed, 95% CI)	0.28 [0.06, 0.50]
23.2 PCI	1	102	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.16 [‐0.55, 0.23]
23.3 CA plus PCI (mixed reporting)	2	557	Std. Mean Difference (IV, Fixed, 95% CI)	0.21 [0.04, 0.38]
24 Length of hospital stay	10	2798	Mean Difference (IV, Random, 95% CI)	‐1.06 [‐1.49, ‐0.63]
25 Length of hospital stay (sensitivity analysis)	3	952	Mean Difference (IV, Random, 95% CI)	‐0.15 [‐0.41, 0.11]
26 Participant satisfaction	1	7021	Risk Ratio (M‐H, Fixed, 95% CI)	1.58 [1.52, 1.63]

Methods	Parallel‐group RCT comparing TRA versus TFA for CA and intervention in people aged ≥ 75 years.
Participants	307 people undergoing CA and intervention All participants referred for inpatient invasive angiography because of suspected coronary artery disease or suspected progression of known coronary disease and aged ≥ 75 years. Inclusion criteria: both the TR and TF approach needed to be clinically possible (e.g. negative Allen's test as assessed by pulse oxymetry), availability for a follow‐up visit 24 h after procedure, normal platelet count and plasmatic coagulation, Hb ≥ 9.0 g/dL, and informed consent to study participation. Exclusion criteria: cardiogenic shock, reduced renal function (creatinine > 1.5 mg/dL, because of potential future need of the radial artery to create an AV fistula) and planned simultaneous right and left heart catheterisation.
Interventions	Group 1: TRA (n = 152) Group 2: TFA (n = 155)
Outcomes	Primary outcomes: procedural success (completion of all diagnostic acquisitions and the percutaneous revascularisation procedure via the TR sheath) occurrence of major complications: death, periprocedural infarction (rise of enzymes to > 3 × upper limit of normal), stroke, or vascular access complications (bleeding, aneurysms, AV fistulae) which required surgical intervention, red cell transfusion (at discretion of treating physician), or which led to a drop in Hb concentration > g/dL Secondary outcomes: minor complications: clinically apparent bleeding of access site which was symptomatic, but did not lead to surgical intervention, transfusion, or a drop in Hb level > 3 g/dL procedural duration (time from the first attempt at arterial puncture to the first cine angiography acquisition and time to completion of diagnostic catheterisation) catheters use (diagnostic and guiding) contrast agent use Study also reported: fluoroscopy time DAP
Notes	Germany Funding source: grant sponsor: Deutsche Stiftung für Herzforschung, Frankfurt, Germany; Grant number: F/05/03
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted.

Methods	Single‐centre parallel‐group RCT. Participants were randomised to TRA group and 437 participants were randomised to TFA group. Randomisation carried out using a concealed computer‐generated random sequence on an intention‐to‐treat basis. Randomisation list managed by nursing staff who informed interventional cardiologist of assigned approach just before procedure. Person responsible for participant registration and randomisation was not involved in treatment of participant.
Participants	836 participants with CA with suspicion of coronary artery disease Exclusion criteria: cardiogenic shock, acute STEMI, history of coronary artery bypass surgery, previous ipsilateral TRA, Raynaud's syndrome, simultaneous right heart catheterisation, necessity for a preprocedural implantation of a transient pacemaker, chronic renal insufficiency (creatinine > 2.0 mg/dL) with the potential necessity of using the radial artery as a native fistula in the future, people having haemodialysis with an AV fistula, absence of an experienced operator, participant refusal, and aged > 75 years because there may be some difficulties with the use of VAS among elderly people
Interventions	Group 1: TRA group (n = 408) Group 2: TFA group (n = 428)
Outcomes	Participants' levels of pain using VAS scale. VAS assessments performed before discharge in people who only underwent CA and 24 h later in people who underwent PCI. All participants were asked to come back 1 month later for the re‐evaluation of pain level according to same scale. Sheath introduction time Total procedural time Total amount of contrast used Mobilisation time Hospital length of stay Access site cross‐overs Use of FFR or IVUS Procedural failure of CA or PCI Use of aspirin, clopidogrel, tirofiban Complications such as catheter‐induced spasm, haematoma (> 5 cm), pseudoaneurysm, bleeding, AV fistula, transient ischaemic attack, cerebrovascular event, and mortality.
Notes	Turkey Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomization was carried out using a concealed computer generated random sequence."
Allocation concealment (selection bias)	Low risk	"The randomisation list was managed by the nursing staff who informed the interventional cardiologist of the assigned approach just before the procedure. The person responsible for patient registration and randomisation was not in any way involved in the treatment of the patient."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes were unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention. We judged risk of bias as low regarding the following outcomes due to being objective and unlikely to be influenced by lack of blinding: sheath introduction time total procedural time total amount of contrast used mobilisation time hospital length of stay use of FFR or IVUS procedural failure of CA or PCI use of aspirin, clopidogrel, tirofiban complications
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Blinding not reported; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participants' levels of pain due to being subjective and unlikely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted

Methods	Multi‐centre multi‐arm RCT. Participants randomly assigned by telephone from a central office to either femoral, brachial, or radial approach.
Participants	People scheduled to undergo coronary angioplasty as treatment of symptomatic ischaemic heart disease with an angiographic significant coronary artery stenosis (primary or restenotic) in a native artery measuring 3.0 mm (balloon diameter as reference) were eligible for the study. 150 men enrolled at 6 different sites between October 1994 and November 1995. Inclusion criteria: men, normal modified Allen's test, good pulsating radial artery, no multiple previous brachial cutdown procedures, no aortobifemoral bypass, no documented subclavian stenosis, and no contraindications to anticoagulant or antiplatelet therapy or to coronary stenting (important proximal coronary artery tortuosity, large side branch originating in lesion, lesion > 15 mm of length, angiographic evidence of thrombus) Exclusion criteria: women because of small size of radial arteries
Interventions	Group 1: brachial approach (n = 38) Group 2: TFA (n = 56) Group 3: TRA (n = 56) Transbrachial arm excluded in our analysis
Outcomes	Primary outcomes: entry site complications (bleeding, haematoma, need for blood transfusion, vascular injury, thrombotic occlusion, need for vascular surgery, nerve damage) mean hospitalisation time post‐stent implantation Secondary outcomes: success rate stent thrombosis Q‐wave MI non‐Q‐wave MI repeated angioplasty coronary bypass surgery cerebrovascular accident haemorrhage death
Notes	Belgium Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients were randomly assigned by telephone from a central office to either femoral or brachial or radial approach."
Allocation concealment (selection bias)	Low risk	"Patients were randomly assigned by telephone from a central office to either femoral or brachial or radial approach."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes are unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	High risk	Study performed "per‐protocol analysis." "Of these 150 patients, nine were excluded for further analysis because stent implantation did not occur using the scheduled approach."
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted.

Methods	Randomised, national, multi‐centre, parallel‐group trial. Operators performed randomisation with personal password through computerised web system.
Participants	707 people with STEMI Inclusion criteria: people admitted to hospital with acute STEMI, within 12 h of symptom onset, and referred for an invasive approach with ability to use both access sites. Exclusion criteria: cardiogenic shock or inability to obtain written informed consent, prior aortobifemoral bypass, absence of bilateral radial or femoral artery pulses, participation in another ongoing clinical trial, negative Allen's test or Barbeau test type D curve, and treatment with oral anticoagulants
Interventions	Group 1: TRA (n = 348) Group 2: TFA (n = 359)
Outcomes	Primary outcomes: cumulative incidence of major bleeding and vascular access site complications requiring intervention at 30 days. NACE defined as a composite of death, MI, stroke, and major bleeding/vascular complications Secondary outcomes: major adverse cardiovascular events (defined as a composite of death, MI, and stroke) technical success access site failure procedural and fluoroscopy times contrast utilisation intensive care unit stay all‐cause mortality at 6 months
Notes	Czech Republic Funding source: supported by Czech Republic Ministry of Health for conceptual development of research organization 00669806, Faculty Hospital in Pilsen, Czech Republic, and Charles University Research Fund (project no. P36). Data analysis supported by International Chair on Interventional Cardiology and Transradial Approach, established at Laval University, Quebec City, Quebec, Canada.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The operators performed randomisation with personal password through computerized web system."
Allocation concealment (selection bias)	Low risk	"The operators performed randomisation with personal password through computerized web system."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	Low risk	Study protocol available and all study's prespecified (primary and secondary) outcomes of interest in review were reported in the prespecified way (ClinicalTrials.gov Identifier: NCT1136187).
Other bias	Low risk	None noted.

Methods	Parallel‐group RCT comparing radial and femoral approaches to PCI using abciximab and 5F guiding‐catheters in AMI
Participants	114 participants consecutively and prospectively enrolled between January 2004 and September 2005 Inclusion criteria: ACS with ST‐segment elevation associated with sustained chest pain Exclusion criteria: haemodynamic instability (i.e. Killip state > 2 or cardiogenic shock), need for an intra‐aortic balloon pump or temporary pacemaker, history of CABG or intolerance to abciximab
Interventions	Group 1: TRA (n = 57) Group 2: TFA (n = 57)
Outcomes	Peripheral arterial complication rates: TIMI major bleeding involved Hb drop > 50 g/L or intracranial haemorrhage or cardiac tamponade TIMI minor bleeding involved Hb drop > 30 g/L but < 50 g/L, with bleeding from a known site or spontaneous gross haematuria, haemoptysis, or haematemesis. Durations of supine bed rest and hospitalisation PCI efficiency and tolerance of the procedure: PCI success rates fluoroscopy duration procedural duration (from peripheral arterial puncture to the end of PCI, excluding sheath removal) number of guiding‐catheters or other devices used during PCI volume of contrast medium used during PCI and kidney tolerance (defined by variations in plasma creatinine concentration) deaths clinical and ECG evidence of sustained or recurrent ischaemia after PCI
Notes	France Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to the nature of the intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals
Selective reporting (reporting bias)	High risk	Study protocol unavailable, but we judged risk of bias as high due to failure to report key outcomes expected to be reported for such a study (adverse cardiac events, i.e. cardiac death, stroke, MI, target lesion revascularisation)
Other bias	Low risk	None noted.

Methods	Single‐centre, parallel‐group RCT. Eligible participants randomly assigned by computer generation (in 2 blocks in a 1:1 ratio) to TFA or TRA
Participants	1024 participants undergoing coronary catheterisation Inclusion criteria: participants referred for diagnostic or interventional cardiac catheterisation were screened for participation. Exclusion criteria: history of coronary artery bypass surgery, cardiogenic shock, known difficulties with the femoral approach (i.e. Leriche syndrome, severe peripheral artery disease, large abdominal aortic aneurysm) or radial approach (i.e. Raynaud's syndrome), simultaneous right heart catheterisation, pathological Allen's test, necessity for a preprocedural implantation of a transient pacemaker, chronic renal insufficiency (creatinine > 2.0 mg/dL) with the potential necessity of using the radial artery as a native fistula in the future, People receiving haemodialysis with an AV fistula, absence of an experienced operator, or participant refusal.
Interventions	Group 1: TFA (n = 512) Group 2: TRA (n = 512)
Outcomes	Procedural success rate Causes for switch to alternative access Procedural duration (time of entry of the participant into the catheterisation laboratory to end of procedure) Fluoroscopy time DAP Amount of contrast agent Vascular access site complications during hospitalisation (pseudoaneurysm, AV fistula, retroperitoneal haematoma, limb ischaemia, surgical vascular repair, loss of radial artery pulse and major bleedings defined as Hb level decline of ≥ 3 g/dL or administration of blood transfusion) Cerebrovascular accident Death Reinfarction or repeat PCI
Notes	Germany Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Eligible patients were randomly assigned by computer generation (in 2 blocks in a 1:1 ratio) to either transfemoral or transradial catheterisation."
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted

Methods	Parallel‐group randomised, multi‐centre, pilot study. Randomisation performed in a concealed manner using sealed envelopes
Participants	50 people with STEMI who were referred for primary or rescue PCI at participating PCI centres were screened for eligibility. For primary PCI, participants could be enrolled within 12 h of symptom onset and for rescue PCI, participants could be enrolled within 12 h of thrombolysis. Rescue PCI was performed for suspected failed reperfusion or reocclusion based on symptoms and ECG changes Exclusion criteria: cardiogenic shock, abnormal Allen's test result, or contraindications to GP IIb/IIIa inhibitor use (active bleeding, major surgery/biopsy/significant trauma in past 6 weeks, systolic blood pressure > 200 mmHg or diastolic blood pressure > 110 mmHg, INR > 2, recent non‐compressible vascular puncture, central nervous system structural damage or stroke/transient ischaemic attack within the last 6 months, baseline platelet count < 100,000 cells/µL)
Interventions	Group 1: TRA (n = 25) Group 2: TFA (n = 25)
Outcomes	Primary outcomes: reperfusion time (time from local anaesthesia infiltration to the first balloon inflation) major bleeding (intracranial or retroperitoneal bleeding, a drop in Hb level > 5 g/dL or haematocrit ≥ 15%, or whole blood or packed red cell transfusions) access site complications (haematoma > 5 cm, pseudoaneurysm, AV fistula, access site rebleeding after initial haemostasis) Secondary outcomes: death reinfarction (new pathological Q waves or re elevation of CK MB higher than the upper limit of normal) urgent revascularisation stroke new congestive heart failure QoL: self‐administered questionnaires were used with the Medical Outcomes Study short form 36‐item health status questionnaire (SF‐36 acute) and a procedure specific VAS rating pain, discomfort, and difficulty walking on a scale from 0 to 10 All outcomes evaluated during initial hospitalisation and at 30 days, except QoL, which was assessed at 24 h and 1 week after PCI.
Notes	Canada Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Low risk	"Randomization was performed in a concealed manner using sealed envelopes."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low regarding the following outcomes due to being objective and unlikely to be influenced by lack of blinding: reperfusion time major bleeding access site complications death reinfarction urgent revascularisation stroke new congestive heart failure
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding QoL due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	High risk	QoL after 1 week was assessed only for 36 participants (28% dropouts) without clearly accounting for the dropouts.
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted

Methods	Parallel‐group RCT. Participants randomly assigned by coin toss to TFA or TRA
Participants	200 participants referred for diagnostic cardiac catheterisation were screened for participation Inclusion criteria: palpable femoral and radial pulses, and normal Allen's test Exclusion criteria: known or suspected vascular disease precluding access, unstable coronary symptoms, need for additional procedures during same hospitalisation, or unable or unwilling to give informed consent Participants not excluded for age, sex, body size, or race
Interventions	Group 1: TRA (n = 101) Group 2: TFA (n = 99)
Outcomes	Primary outcomes: participant preference QoL assessed in all participants before catheterisation and 1 day and 1 week thereafter cost Secondary outcomes: procedural success: defined as the ability to complete all planned components from the assigned access site major vascular complications: defined as pseudoaneurysm formation, blood transfusion, or limb ischaemia resulting in need for vascular surgery haematoma: defined as any measurable discolouration at the access site cross‐over to alternative access sheath insertion time catheterisation time haemostasis time total procedure duration fluoroscopy time contrast volume catheters per case hospital stay hospital readmission Study also reported: stroke percutaneous coronary revascularisation All outcomes assessed 1 day and 1 week after catheterisation, QoL assessed before catheterisation
Notes	USA Funding source: supported by a grant from Cordis Corporation. Dr Cohen was supported in part by a Clinician‐Scientist Award from the American Heart Association.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients were randomly assigned by coin toss to either transfemoral or transradial catheterisation."
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low regarding the following outcomes due to being objective and unlikely to be influenced by lack of blinding: procedural success major vascular complications haematoma cross‐over to alternative access sheath insertion time catheterisation time haemostasis time total procedure duration fluoroscopy time contrast volume catheters per case hospital stay hospital readmission cost
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding of assessors of participant preference and QoL; however, blinding impossible due to nature of intervention, but we judged risk of bias as high due to these outcomes being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	"No patient was lost to follow‐up."
Selective reporting (reporting bias)	High risk	Study protocol unavailable, but we judged risk of bias as high due to failure to report key outcomes expected to be reported for such a study (adverse cardiac events, i.e. cardiac death, MI)
Other bias	Low risk	None noted

Methods	Parallel‐group, national, multi‐centre, non‐inferiority RCT comparing the radial versus femoral approach using VCD. Randomisation using a randomised sequence obtained by computer algorithms and maintained in individual, opaque, and closed envelopes to conceal the allocation process.
Participants	Between July 2012 and March 2015, 240 people with NSTEMI undergoing early invasive strategy were randomised to radial or femoral approach with use of the Angio‐Seal VCD (St Jude Medical, St Paul, MN, USA). Participants had to have at least 2 of 3 high‐risk criteria: ischaemic changes in 12‐lead ECG, positive biomarkers, or age > 60 years. Inclusion criteria: non‐ST‐segment elevation ACS intention to submit participant to an early invasive strategy consisting of CA immediately followed by PCI, when applicable, in first 72 h after admission participant informed about nature of study and agreeing with general terms and provided informed consent, as approved by the Research Ethics Committee of the respective centre participant eligible for CA and both radial and femoral PCI with the following prerequisites: palpable radial artery with the Allen's or normal oximetry tests, familiarity of operator with radial and femoral techniques using AngioSeal, agreement of operator to use access route determined by randomisation process Exclusion criteria: aged < 18 years pregnancy chronic use of vitamin K antagonists, direct thrombin inhibitors, or oral factor Xa antagonists hypersensitivity to antiplatelet or anticoagulant (or both) drugs active bleeding or high bleeding risk (severe liver failure, active peptic ulcer, creatinine clearance < 30 mL/min, platelets count < 100,000 mm³) uncontrolled systemic hypertension cardiogenic shock previous myocardial revascularisation surgery with ≥ 1 internal mammary or radial artery graft documented chronic peripheral arterial insufficiency preventing use of femoral technique severe concomitant disease with life expectancy < 12 months participation in drug or device investigative clinical trials in last 30 days indication of elective PCI to be performed at a moment different from immediately after CA medical, geographic, or social conditions impairing participation in study or inability to understand and sign informed consent term
Interventions	Group 1: TRA (n = 120) Group 2: TFA (n = 120)
Outcomes	Primary outcome: assessed at 30 days Vascular complications at arterial puncture site 30 days after procedure (major bleeding, retroperitoneal haematoma, compartment syndrome, haematoma ≥ 5 cm, pseudoaneurysm, AV fistula, infection, limb ischaemia, arterial occlusion including asymptomatic occlusion, adjacent nerve injury, or need for vascular surgical repair) Secondary outcomes: assessed at 12 months Individual components of primary outcome Haematoma < 5 cm Major bleeding unrelated to puncture site or myocardial revascularisation surgery Device success and cross‐over rate Cardiovascular death, AMI, or stroke Time to discharge
Notes	Brazil Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients will be randomised for the radial or femoral technique with VCD using a randomised sequence obtained by computer algorithms and maintained in individual, opaque and closed envelopes to conceal the allocation process."
Allocation concealment (selection bias)	Low risk	"Patients will be randomised for the radial or femoral technique with VCD using a randomised sequence obtained by computer algorithms and maintained in individual, opaque and closed envelopes to conceal the allocation process."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes are unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	Low risk	Study protocol available and all study's prespecified primary outcomes (30 days) of interest in review were reported in the prespecified way. **However, it must be noted that none of the study's prespecified secondary outcomes (1 year of follow‐up) were reported in this publication or in any other separate following publications.
Other bias	Low risk	None noted

Study	Reason for exclusion
Bhat 2017	Quasi‐randomised study "with odd serials going into one arm and even into another."
Chodor 2009	Quasi‐randomised study "randomisation was conducted in the admission room based on year of birth: group I included individuals born in even years and group II those born in odd years."
Chodor 2011	Quasi‐randomised study "randomisation was based on year of birth (even number‐TRA; odd number‐TFA)."
Genereux 2011	Participants were randomised to interventions other than those of the review (heparin + GPIIb/IIIa vs bivalirudin) and comparison of TRA vs TFA was a post‐hoc analysis.
Kallinikou 2016	Non‐randomised study "allocation by the judgement of the clinician."
Marti 2015	Participants were randomised to interventions other than those of interest to the review.
Qi 2017	Non‐randomised study "allocation by the judgement of the clinician."
Scalone 2014	Participants were receiving heart transplants undergoing routine angiogram; they did not meet inclusion criteria of review.

Methods	Single‐centre parallel‐group RCT
Participants	360 participants (200 men, 160 women) single centre, aged 25‐78 years, mean (± SD) age 59.3 ± 8.9 years Inclusion criteria: people with diagnostic CA or coronary artery intervention
Interventions	Group 1: TRA (n = 180) Group 2: TFA (n = 180)
Outcomes	Success of intervention Duration of puncture (min) Duration of operation (min) Duration of hospitalisation (days) Complications
Notes	China Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	High risk	Study protocol unavailable, but we judged risk of bias as high due to failure to report key outcomes expected to be reported for such a study (adverse cardiac events, i.e. cardiac death, stroke, MI, target lesion revascularisation)
Other bias	Low risk	None noted

Methods	Single‐centre parallel‐group RCT where participants were randomly divided into TRA group and TFA group.
Participants	From August 2005 to September 2008, 200 people with AMI were included in study which took place in Department of Cardiology, The Tenth People's Hospital, Tongji University, Shanghai, China. After written informed consent obtained, participants were randomly divided into TRA group and TFA group. Exclusion criteria were clinical indications to femoral approach due to cardiogenic shock, history of coronary bypass graft, negative Allen's test, and non‐palpable radial artery.
Interventions	Group 1: TRA (n = 100) Group 2: TFA (n = 100)
Outcomes	Puncture time. Cannulation time: defined as the time from participant arrival at the catheterisation laboratory to the effective placement of the arterial sheath. Reperfusion time: defined as the time from the cannulation to balloon inflation. Procedure time: defined as the time from the first attempt to puncture the artery to the end of the angioplasty. Fluoroscopy time. Procedure success: defined as residual diameter stenosis < 30% with grade 3 coronary flow according to the classification of the TIMI trial. Access site bleeding: defined as major if Hb loss ≥ 2 mmol/L, administration of blood transfusions, and needing vascular repair, and as minor as haematoma formation not requiring specific therapy MACE: defined as death, recurrent MI, or TVR Hospital stay Endpoints recorded from start of procedure to 1‐month follow‐up
Notes	China Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	"One month follow‐up was complete in all patients."
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted

Methods	Multi‐centre, parallel‐group RCT. Participants were randomly assigned (1:1) to radial or femoral access by a 24‐h computerised central automated voice response system located at the Population Health Research Institute in Hamilton, Canada. Randomisation in permuted blocks of variable sizes (2, 4, and 6), stratified by centre.
Participants	7021 people with ACS (with or without ST elevation) randomised to radial or femoral access for CA/intervention Inclusion criteria: eligible for RIVAL (RadIal Vs femorAL access for coronary intervention) if they presented with non‐ST‐segment elevation ACS or ST‐segment elevation ACS, they were to be managed with an invasive approach, they had intact dual circulation of the hand documented by Allen's test Exclusion criteria: aged < 18 years; active bleeding or significant increased risk of bleeding (severe hepatic insufficiency, current peptic ulceration, proliferative diabetic retinopathy); uncontrolled hypertension; cardiogenic shock; prior CABG surgery with use of > 1 internal mammary artery; documented severe peripheral vascular disease precluding a femoral approach; previously entered in study; investigational treatment (drug or device) within the previous 30 days; medical, geographic, or social factors making study participation impractical or inability to provide written informed consent and to understand the full meaning of the informed consent
Interventions	Group 1: TRA (n = 3507) Group 2: TFA (n = 3514)
Outcomes	Primary outcome: composite of death, MI, stroke or non‐CABG‐related major bleeding up to 30 days Secondary outcomes at 48 h and 30 days: death, MI, or stroke non‐CABG related major bleeding Other outcomes: components of primary outcome at 48 h and 30 days PCI procedural success access site cross‐over major vascular access site complications at 48 h and 30 days after the procedure (pseudoaneurysms requiring ultrasound compression, thrombin injection, or surgical repair; and large haematomas requiring prolonged hospitalisation, AV fistulae, limb ischaemia, or damage to adjacent nerve) minor bleeding Safety outcomes: non‐CABG TIMI major bleeding CABG‐related bleeding non‐CABG‐related blood transfusions all blood transfusions Radiation outcomes: radiation measurements in the form of air kerma and DAP collected in 2569 participants enrolled from sites that had facilities to provide these measurements fluoroscopy time PCI complications: abrupt closure no reflow dissection with reduced flow coronary perforation catheter thrombus stent thrombosis Study also reported: PCI procedural time PCI contrast volume length of stay in hospital persistent pain at access site for ≥ 2 weeks participant prefers TRA next procedure
Notes	Canada, Finland, India, Czech Republic, Poland, Spain, Israel, Brazil, France, and England Funding source: funding for the RIVAL trial was provided by Sanofi‐Aventis, Population Health Research Institute, and the Canadian Network and Center for Trials Internationally (CANNeCTIN), which was funded by the Canadian Institutes of Health Research.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients were randomly assigned (1:1) to radial or femoral access by a 24‐hour computerised central automated voice response system located at the Population Health Research Institute in Hamilton, Canada."
Allocation concealment (selection bias)	Low risk	"Patients were randomly assigned (1:1) to radial or femoral access by a 24‐hour computerised central automated voice response system located at the Population Health Research Institute in Hamilton, Canada."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Participants and investigators were not masked to treatment allocation, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	"A masked central committee adjudicated the primary outcome, components of the primary outcome, and stent thrombosis."
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding of assessors of outcomes of pain at access site and participant preference for next procedure; however, blinding impossible due to nature of intervention, but we judged risk of bias as high due to these outcomes being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	Low risk	Study protocol available and all study's prespecified (primary and secondary) outcomes of interest in review were reported in prespecified way (ClinicalTrials.gov Identifier: NCT01014273).
Other bias	Low risk	None noted

Methods	Single‐centre multi‐arm RCT. Randomisation performed using sealed envelopes containing a code for TR, transbrachial, or TF angioplasty. Envelopes were ordered at random and included a registration number from 1 to 900.
Participants	900 included who underwent PTCA at the department. Inclusion criteria: stable and unstable angina selected for single or multi‐vessel PTCA of lesions in native coronary arteries and venous bypass grafts. Underwent PTCA by the TRI, transbrachial approach, or TFA. Exclusion criteria: vascular status (absence of pulse in femoral, brachial, or radial arteries; abnormal Allen's test results; failed previous arterial access); cardiac status (chronic total occlusion, acute MI, expected severe haemodynamic deterioration during balloon inflation or after PTCA failure leading to intra‐aortic pumping or right heart catheterisation for haemodynamic monitoring, expected need for a temporary pacemaker); procedural (ad hoc PTCA after TF diagnostic catheterisation, indwelling sheath from previous arterial puncture, planned primary stent implantation, planned coronary atherectomy); no consent
Interventions	Arm 1: TRA PTCA (n = 300) Arm 2: transbrachial approach PTCA (n = 300) Arm 3: TFA PTCA (n = 300) The transbrachial arm was excluded in our analysis.
Outcomes	Primary outcomes: successful puncture (access site related) major access site complication (access site related) residual stenosis (> 50%) MACE: death, MI, CABG, or repeat PTCA Secondary outcomes: change in minimal lumen diameter and % diameter stenosis: assessed by quantitative coronary analysis procedural time fluoroscopy time consumption of angioplasty equipment hospital stay radial artery occlusion at hospital discharge major bleeding Primary outcomes recorded from start of procedure to 1‐month follow‐up.
Notes	The Netherlands Funding source: study supported by a grant from Scimed Life Systems, Inc., Maple Grove, MN.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation.
Allocation concealment (selection bias)	Low risk	"Randomization was performed by opening a sealed envelope containing a code for either transradial (R), transbrachial (B) or transfemoral (F) angioplasty."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted.

Methods	Parallel‐group, open‐label RCT. After informed consent acquired, participants randomised to femoral or radial access. The random sequence generated and allocation concealed using opaque envelopes containing icon‐cards.
Participants	103 people with STEMI screened between September 2010 and October 2012. Inclusion criteria: pain duration between 20 min and 24 h, ST‐segment elevation measured at J point in 2 contiguous leads ≥ 0.25 mV in men aged < 40 years, ≥ 0.2 mV in men aged > 40 years, or ≥ 0.15 mV in women in leads V2‐V3 or ≥ 0.1 mV in other leads (or both) (in the absence of left ventricular hypertrophy or LBBB) or newly emerged LBBB, aged ≥ 18 years, and person's informed consent Exclusion criteria: INR > 1.4, thrombocytopenia < 100 × 10³, previous CABG, known vascular access difficulties or complications, active bleeding, gastric or duodenal peptic ulcer, current or planned dialysis, severe liver failure (MELD > 10 points), uncontrolled hypertension (> 160/100 mmHg), cardiogenic shock, and low compliance to long‐term follow‐up.
Interventions	Group 1: TRA (n = 52) Group 2: TFA (n = 51)
Outcomes	Primary outcome: composite of major bleeding (according to the definition published in the Randomised Evaluation in Percutaneous Coronary Intervention Linking Angiomax to Reduced Clinical Events 2 (REPLACE‐2) study) and minor bleedings defined by the EASY scale (radial arm ‐ TR) and FEMORAL scale (femoral access ‐ TF) Secondary outcomes: angiographic success: defined by TIMI flow classification as TIMI 3 and residual stenosis < 20% long‐term mortality clinical efficacy defined as survival and no MACE during in‐hospital stay: stroke death sudden cardiac arrest major bleeding REPLACE‐2 repeat revascularisation Study also reported: cost length of hospitalisation QoL before PCI using EQ‐5D‐3L VAS, and 2 h and 4 days after PCI using EQ‐5D‐3L health utility score
Notes	Poland Funding source: investigator‐initiated trial funded by authors.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Low risk	"The random sequence was generated and the allocation was concealed using opaque envelopes containing icon‐cards."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Open‐label RCT, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Open‐label RCT; however, blinding impossible due to nature of intervention, but we judged risk of bias as low regarding the following outcomes due to being objective and unlikely to be influenced by lack of blinding: composite of major bleeding and minor bleedings angiographic success long‐term mortality clinical efficacy cost length of hospitalisation
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Open‐label RCT; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding QoL due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted

Methods	Parallel‐group RCT. After informed consent was obtained, catheterisation procedure was performed randomly by TFA or TRA by same operator.
Participants	After completion of procedure, only cases were included for data analysis who fulfilled the following criteria: procedure was uncomplicated and right femoral or right radial artery could be accessed without difficulties; only coronary angiograms were performed (i.e. cases with bypass graft angiography, left ventricular cineangiography, or aortography were excluded); coronary interventions were done electively on 1 single epicardial vessel. Cases with bypass graft intervention, bifurcation lesions, and chronic total occlusions were excluded.
Interventions	Group 1: TRA (n = 146) Group 2: TFA (n = 151)
Outcomes	Fluoroscopy time DAP (Gy/cm²) Radiation exposure (mSv)
Notes	Germany Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	High risk	"Only cases were included for data analysis who fulfilled the following criteria: the procedure was uncomplicated and the right femoral or right radial artery could be accessed without difficulties." Study performed "per‐protocol analysis."
Selective reporting (reporting bias)	High risk	Study protocol unavailable, but we judged risk of bias as high due to failure to report key outcomes expected to be reported for such a study (adverse cardiac events (i.e. cardiac death, stroke, MI, target lesion revascularisation).
Other bias	Low risk	None noted

Methods	Parallel‐group RCT of TRA versus TFA
Participants	From June 2004 to June 2006, 370 consecutive participants admitted to hospital diagnosed as AMI within 12 h from onset of chest pain. Exclusion criteria: for TRA: negative Allen's test, aorto‐arteritis, cardiogenic shock, non‐palpable radial artery, severe tortuosity of radial arteries or body height < 150 cm. 244 men and 126 women randomly allocated to TRA or TFA
Interventions	Group 1: TRA (n = 184) Group 2: TFA (n = 186)
Outcomes	Cannulation time Total procedure time Recanalisation with TIMI III flow of culprit vessels Stent implanted Successful rates of puncture Vascular access site complications: local hematoma AV fistula pseudoaneurysm angiospasm vagus reflex difficult micturating low back pain
Notes	China Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low regarding the following outcomes due to being objective and unlikely to be influenced by lack of blinding: cannulation time total procedure time recanalisation with TIMI III flow of culprit vessels stent implanted successful rates of puncture vascular access site complications
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding low back pain due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	High risk	Study protocol unavailable, but we judged risk of bias as high due to failure to report key outcomes expected to be reported for such a study (adverse cardiac events (i.e. cardiac death, stroke, MI, target lesion revascularisation)
Other bias	Low risk	None noted

Methods	Multi‐arm RCT. Learning curve taken into account in randomisation of participants. 1 operators with experience in left radial approach randomised participants into left TRA and TFA and the other operator, with previous right radial approach experience into right TRA and TFA. Randomisation ratio was 2 participants in TRA group for 1 in TFA group to obtain 3 even groups.
Participants	210 participants randomised over 6‐month period between November 1998 and April 1999. Inclusion criteria: normal Allen's test and participant informed consent. Exclusion criteria: clinical (AMI), angiographic (previous bypass grafting), or technical (known difficulty with TFA, right heart catheterisation, simultaneous renal or aortic angiography, and absence of indication for ventricular angiogram). No other selection made as to symptoms, age, gender, weight, or height.
Interventions	Group 1: TFA (n = 70) Group 2: right TRA (n = 70) Group 3: left TRA (n = 70) Right and left TRA groups were combined in our analysis.
Outcomes	Procedural success Procedural time Radiation time Angiographic quality Complications Bed confinement Duration of hospital stay Cost‐effectiveness Participant comfort Cerebrovascular accident
Notes	France Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low regarding the following outcomes due to being objective and unlikely to be influenced by lack of blinding: procedural success procedural time radiation time angiographic quality complications bed confinement duration of hospital stay cost‐effectiveness cerebrovascular accident
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant comfort due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	High risk	Study protocol unavailable, but we judged risk of bias as high due to failure to report key outcomes expected to be reported for such a study (adverse cardiac events (i.e. cardiac death, stroke, MI, target lesion revascularisation).
Other bias	Low risk	None noted

Methods	Multi‐centre parallel‐group RCT. Randomisation carried out using blinded allocation list before clinical examination of femoral arteries or assessment of hand blood supply.
Participants	Aged > 80 years presenting for CA or PCI. Inclusion criteria: informed consent. All clinical presentations were permissible (including ST‐segment elevation AMI), except cardiogenic shock. Exclusion criteria: history of bypass grafting surgery using 2 in situ internal mammary arteries, documented severe peripheral artery disease, or with previous unsuccessful approaches
Interventions	Group 1: TRA (n = 192) Group 2: TFA (n = 185)
Outcomes	Primary outcome: Composite vascular endpoint comprising ≥ 1 of the following: vascular complications requiring surgical intervention blood transfusion reduction in Hb by ≥ 3 g/dL or haematocrit by 10% acute arm or leg ischaemia forearm compartment syndrome other vascular complications resulting in delayed discharge from hospital Secondary outcomes: vascular complications included in the combined endpoint cholesterol embolisation cerebrovascular accidents occlusion of the radial artery false aneurysm large haematoma (resulting in delayed hospital discharge) haematoma procedural efficacy according to approach: percentage crossover to a different approach number of catheters used total procedure duration X‐ray exposure time volume of contrast media used primary success (all lesions successfully treated without complications) angioplasty‐related cardiac complications Other outcomes: non‐Q‐wave MI Q‐wave MI CABG death
Notes	France and UK Funding source: not described We contacted the authors for missing data regarding randomisation methods, but received no response.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted.

Methods	Parallel‐group RCT. Participants were randomised to either radial or femoral group access prior to the catheterisation procedure.
Participants	142 people with ACS admitted to hospital between April and July 1997. Participants within this group who underwent coronary stenting were included.
Interventions	Group 1: TRA (n = 74) Group 2: TFA (n = 68)
Outcomes	Procedural success: defined as a reduction of the target lesion to < 25% luminal diameter using visual angiographic assessment without complication Death/CABG/MI Access site complications: defined as a bleeding vascular complication that prolonged hospitalisation Procedural MI: defined as an abnormal CPK elevation following the procedure Postprocedure length of stay Total hospital length of stay Total hospital charge
Notes	USA Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	High risk	Study performed "as treated" analysis: (during course of study, 6/74 (8%) participants randomised to TRA had a negative Allen's test or Doppler examination (or both) suggesting an incomplete palmer arch; these were included in TFA group. Of the remaining 68 participants in TRA group, 65 participants had their procedures performed TR. In 3 participants, radial artery was not successfully cannulated, and these were also included in TFA group. Thus, 77 participants who had their procedures performed from TFA included 9 radial cross‐overs).
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted

PERMALINK

Transradial versus transfemoral approach for diagnostic coronary angiography and percutaneous coronary intervention in people with coronary artery disease

Ahmed A Kolkailah

Rabah S Alreshq

Ahmed M Muhammed

Mohamed E Zahran

Marwah Anas El‐Wegoud

Ashraf F Nabhan

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Summary of findings for the main comparison. Transradial compared to transfemoral approach for diagnostic coronary angiography and percutaneous coronary intervention in people with coronary artery disease.

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Timing of outcome assessment

Search methods for identification of studies

Electronic searches

Searching other resources

Ongoing trials

Checking reference lists

Personal communication

Data collection and analysis

Selection of studies

1.

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Dichotomous data

Continuous data

Unit of analysis issues

Cluster‐randomised trials

Multi‐arm intervention trials

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

GRADE and 'Summary of findings' table

Main outcomes for 'Summary of findings' table

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

Included studies

Design

Setting

Participants

Age

Gender

Clinical characteristics

Intervention

Funding source

Excluded studies

Studies awaiting classification

Ongoing studies

Risk of bias in included studies

2.

3.

Allocation

Blinding

Incomplete outcome data

Methods	Single‐centre, parallel‐group RCT. Participants randomised to TRA or TFA in 1:1 ratio using opaque, numbered, sealed envelopes containing randomisation assignment based on a computer‐generated random sequence.
Participants	128 participants having previously undergone CABG surgery and referred for cardiac catheterisation at Dallas VA Medical Center. Inclusion criteria: previously undergone CABG surgery and referred for diagnostic or interventional cardiac catheterisation. Exclusion criteria: presenting with ST‐segment elevation AMI, abnormal Allen's test results, known difficulty obtaining vascular access via either femoral or radial artery, or aged > 90 years.
Interventions	Group 1: TRA (n = 64) Group 2: TFA (n = 64)
Outcomes	Primary outcome: volume of radiographic contrast administered during cardiac catheterisation Secondary outcomes: fluoroscopy time total procedure time (defined as the interval between administration of local anaesthesia for obtaining vascular access and removal of the last catheter) radiation exposure of participant and operators vascular access cross‐over vascular access complications (haematoma, aneurysm, pseudoaneurysm, AV fistula formation, retroperitoneal haematoma, dissection, and limb ischaemia) periprocedural MACE (defined as all‐cause death, Q‐wave MI, recurrent angina requiring urgent repeat TVR with PCI or coronary bypass surgery, and stroke occurring before hospital discharge). participant satisfaction regarding access site and overall discomfort, earliest time to ambulation after procedure, and future access site preference. All endpoints assessed separately for diagnostic angiography and PCI (i.e. measurements for diagnostic catheterisation ended with completion of diagnostic angiography and measurements for PCI began on PCI initiation and ended on PCI completion). Data regarding participant satisfaction collected 1 day after procedure.
Notes	USA Funding source: Dr Michael received a cardiovascular training grant from the National Institutes of Health, Award Number T32HL007360. Dr Banerjee received a research grant from Boston Scientific and an institutional research grant from Gilead.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients were randomised to TR or TF access in a 1:1 ratio using opaque, numbered, sealed envelopes containing randomisation assignment based on a computer generated random sequence."
Allocation concealment (selection bias)	Low risk	"Patients were randomised to TR or TF access in a 1:1 ratio using opaque, numbered, sealed envelopes containing randomisation assignment based on a computer generated random sequence."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low regarding the following outcomes due to being objective and unlikely to be influenced by lack of blinding: volume of radiographic contrast administered during cardiac catheterisation fluoroscopy time total procedure time radiation exposure of participant and operators vascular access cross‐over vascular access complications peri‐procedural MACE.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant satisfaction due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	Low risk	Study protocol available and all study's prespecified (primary and secondary) outcomes of interest in review were reported in prespecified way (ClinicalTrials.gov Identifier: NCT01446263).
Other bias	Low risk	None noted

Methods	Multi‐centre, parallel‐group, open‐label RCT. After informed consent obtained, but before obtaining arterial access for procedure, participants randomly assigned (1:1) to TRA or TFA. Web‐based randomisation performed using a DCRI proprietary Simple Internal Randomization Engine in block fashion within sites. Participants enrolled at sites performing ad hoc PCI randomised before diagnostic angiography. All primary endpoint events were adjudicated by an independent Clinical Events Committee.
Participants	1775 women (691 undergoing PCI) randomised at 60 sites. Inclusion criteria: women, aged ≥ 18 years, undergoing urgent or elective PCI, undergoing diagnostic angiography for ischaemic symptoms with possible PCI, and ability to provide informed consent for trial participation. Exclusion criteria: peripheral arterial disease prohibiting vascular access, bilateral abnormal Barbeau tests, haemodialysis access (AV fistula or graft) in the arm to be used for PCI in case of assignment to radial approach (the opposite arm may be used for radial access if a dialysis graft is present in 1 arm provided that the opposite arm has a normal Barbeau test result), valvular heart disease requiring valve surgery, planned right heart catheterisation, primary PCI for STEMI, presence of bilateral internal mammary artery coronary bypass grafts, participation in investigational drug or device study currently or within 30 days before enrolment, INR ≥ 1.5 while treated with oral vitamin K antagonists (i.e. warfarin), receipt of oral factor Xa or IIa inhibitors ≤ 24 h before procedure, planned staged PCI within 30 days after index procedure.
Interventions	Group 1: TRA (n = 891) Group 2: TFA (n = 884)
Outcomes	Primary outcome: composite of BARC type 2, 3, or 5 bleeding or vascular complications (any of the following that require surgical intervention, including thrombin injection, AV fistula, arterial pseudoaneurysm, or arterial occlusion) occurring from the first arterial access after randomisation through 72 h or hospital discharge, whichever occurs first). procedural failure: defined as inability to complete the procedure from the assigned vascular access site (access site cross‐over) Secondary outcomes: procedure duration participant radiation dose (air kerma) contrast volume composite of death, vascular complications, or unplanned revascularisation at 30 days in participants undergoing PCI participant's access site preference for their next procedure Secondary endpoints were assessed only in the subgroup of participants undergoing PCI.
Notes	England, USA, Canada Funding source: Abbott Vascular, Medtronic Vascular (grant no. A 1054367), Terumo Medical, The Medicines Company, Daiichi Sankyo/Eli Lilly and Company (grant no. H7TUS‐X014), ACIST Medical, Guerbet, FDA Office of Women's Health (grant no. HHSF223201111381P), and The Duke Clinical Research Institute funded this study.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients were randomly assigned to radial or femoral access. The allocation ratio for randomisation is 1:1 between treatment arms, and Web‐based randomisation was performed using a DCRI proprietary Simple Internal Randomization Engine in block fashion within sites."
Allocation concealment (selection bias)	Low risk	"Patients were randomly assigned to radial or femoral access. The allocation ratio for randomisation is 1:1 between treatment arms, and Web‐based randomisation was performed using a DCRI proprietary Simple Internal Randomization Engine in block fashion within sites."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	"All primary end point events will be adjudicated by an independent Clinical Events Committee."
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participants' access site preference for their next procedure due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	"Follow‐up for the primary endpoints was available for 99.3% of the total randomised cohort."
Selective reporting (reporting bias)	Low risk	Study protocol available and all study's prespecified (primary and secondary) outcomes of interest in review were reported in the prespecified way (ClinicalTrials.gov Identifier: NCT01406236).
Other bias	Low risk	None noted.

Methods	Multi‐arm RCT. Participants were randomly assigned to TRA and TFA for cardiac catheterisation.
Participants	75 participants referred for diagnostic cardiac catheterisation were screened for study participation. Inclusion criteria: aged ≥ 18 years, easily palpable femoral and radial pulses, and normal Allen's test. Exclusion criteria: vascular disease of upper or lower extremities precluding access at either femoral or radial artery, prior femoral arterial graft surgery, unstable coronary syndromes, haemodynamically unstable patients with MI who require an intervention within 7 days, people for whom additional procedures were planned at the same setting or during the same hospital stay (e.g. PCI, peripheral angiography or intervention, electrophysiology study), and people unable or unwilling to provide informed consent.
Interventions	Group 1: TFA 6F with AngioSeal closure device (n = 25) Group 2: TFA 4F without a closure device (n = 25) Group 3: TRA (n = 25) The 2 TFA groups were combined in our analysis.
Outcomes	Primary outcome: QoL: assessed by the Medical Outcomes Study Short Form 36‐item health questionnaire (acute SF‐36) and a series of procedure‐specific questions. Participants were evaluated prior to catheterisation and at 1 day and 1 week. Secondary outcomes: procedural success: defined as the ability to complete all planned components from assigned access site without cross‐over to another access management strategy. haematoma: defined as any measurable subcutaneous induration at access site. postprocedural major vascular complications: defined as bleeding requiring blood transfusion, limb ischaemia requiring revascularisation, or formation of pseudoaneurysm requiring treatment. cost sheath insertion time catheterisation time haemostasis time fluoroscopy time contrast amount angiographic quality Assessed at 1 day and 1 week post procedure follow‐up.
Notes	USA Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation.
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low regarding the following outcomes due to being objective and unlikely to be influenced by lack of blinding: procedural success haematoma postprocedural major vascular complications cost sheath insertion time catheterisation time haemostasis time fluoroscopy time contrast amount angiographic quality
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding QoL due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Complete data obtained for 73 participants at 1‐day and 70 participants at 1‐week follow‐up visits.
Selective reporting (reporting bias)	High risk	Study protocol unavailable, but we judged risk of bias as high due to failure to report key outcomes expected to be reported for such a study (adverse cardiac events (i.e. cardiac death, stroke, MI, target lesion revascularisation).
Other bias	Low risk	None noted

Methods	Multi‐centre, parallel‐group RCT. All enrolled participants were randomised (1:1 ratio) to TRA or TFA according to opaque, numbered, sealed envelopes with randomisation based on a computer‐generated random series and stratified by centre.
Participants	People with suspected STEACS planned for early revascularisation strategy (within 24‐h of symptom onset) were eligible. Exclusion criteria: contraindication to either radial or femoral vascular access (e.g. abnormal Allen's test or known severe peripheral vascular disease), recent stroke (within 4 weeks), anticoagulant therapy assumption with an INR 2, or other severe bleeding diathesis. Cardiogenic shock or haemodynamic instability (or both) did not preclude enrolment.
Interventions	Group 1: TRA (n = 500) Group 2: TFA (n = 501)
Outcomes	Primary outcome: NACEs: defined as the composite of cardiac death, MI, stroke, target lesion revascularisation, and non‐CABG‐related bleeding at 30 days. Secondary outcomes: 30‐day individual components of NACEs hospital stay
Notes	Italy and The Netherlands Funding source: no extramural funding used to support the work, and authors were solely responsible for design, conduct, and final contents of study. We contacted the authors regarding long‐term outcome data. They responded that the data were unavailable.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"All enrolled patients were randomised (1:1 ratio) to radial or femoral access according to opaque, numbered, sealed envelopes with randomisation based on a computer generated random series and stratified by centre."
Allocation concealment (selection bias)	Low risk	"All enrolled patients were randomised (1:1 ratio) to radial or femoral access according to opaque, numbered, sealed envelopes with randomisation based on a computer generated random series and stratified by centre."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Participants and investigators were not blinded to procedure, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	"Endpoint adjudication was performed by a blinded central independent clinical‐event committee."
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No participant was lost at 30‐day follow‐up; thus, all 1001 participants were included in final intention‐to‐treat analyses. 14 of participants originally randomised to femoral arm were crossed over to radial access 47 of participants originally randomised to radial arm were crossed over to femoral access; however, intention‐to‐treat analysis performed.
Selective reporting (reporting bias)	Low risk	Study protocol available and all study's prespecified primary outcomes (30‐days) of interest in review were reported in prespecified way (ClinicalTrials.gov Identifier: NCT01420614). **However, it must be noted that none of the study's prespecified secondary outcomes (1‐year follow‐up) were reported in this publication or in any other separate subsequent publications.
Other bias	Low risk	None noted

Methods	Parallel‐group RCT. Participants randomised to TRI or TFI groups before arterial puncture.
Participants	149 participants with AMI Inclusion criteria: AMI within 12 h from onset of study if written informed consent obtained before emergency catheterisation, onset and location of infarction were clearly documented, participant had not received any type of thrombolytic therapies, aged > 20 years, and normal Allen's test. When a participant came into the catheterisation laboratory with shock, that participant was considered eligible for randomisation if the radial artery pulse could be felt. Exclusion criteria: radial artery pulse too weak for successful radial artery puncture, culprit vessel was previous coronary bypass graft, and operator for that particular person did not consider that both TRI and TFI would be equally feasible.
Interventions	Group 1: TRI (n = 77) Group 2: TFI (n = 72)
Outcomes	Primary outcomes: MACE during initial hospitalisation period MACE at 9‐month follow‐up period Secondary outcomes: achievement of final TIMI 3 flow time from onset of AMI to end of PCI procedures restenosis rate Other outcomes: number of guiding catheters used fluoroscopy time total procedure time total amount of dye used number of stents severe bleeding complications cost for first admission length of hospital stay (days) after excluding people with inhospital death
Notes	Japan Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Total dropouts at 9 months' follow‐up: 10/149 (6%); TRI group: 4/77 (5%); TFI group: 6/72 (8%)
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted

Methods	Multi‐arm RCT. Randomised to undergo TFA, right radial approach (RRA), or left radial approach (LRA) using a block design to ensure that same number of participants would be assigned to each technique for arterial access. Randomisation list kept by nursing staff, who informed interventional cardiologist of assigned approach prior to procedure.
Participants	1005 consecutive participants who underwent cardiac catheterisation between January 2007 and July 2007. Allen's test was performed in all participants assigned to radial approach; TR catheterisation not attempted if results of Allen's test were clearly abnormal (in which case, it was considered that the procedure could not be completed via said approach). Exclusion criteria: none
Interventions	Group 1: left TRA (n = 335) Group 2: right TRA (n = 335) Group 3: TFA (n = 335) Right and left TR groups were combined in our analysis.
Outcomes	Primary outcome: success rate in the assigned approach: catheterisation could be performed from start to finish via the initially assigned approach, without applying exclusion criteria. Secondary outcomes: success rate among those cases in which there was no contraindication for the assigned approach or cross‐over due to a decision made by operator or by participant differences in procedure times (cannulation time, time to diagnosis, fluoroscopy time, compression time, time to ambulation, total diagnostic procedure, and interventional procedure) amount of contrast medium administered in the diagnostic procedures incidence of vascular complications
Notes	Spain Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Low risk	"The randomisation list was in the hands of the nursing staff, who informed the interventional cardiologist of the assigned approach prior to the procedure."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	High risk	Study protocol unavailable, but we judged risk of bias as high due to failure to report key outcomes expected to be reported for such a study (adverse cardiac events (i.e. cardiac death, stroke, MI, target lesion revascularisation)
Other bias	Low risk	None noted

Methods	Multi‐centre parallel‐design RCT. Participants presenting with AMI were randomly assigned (1:1) to TRA or TFA group for diagnostic CA and PCI, if indicated.
Participants	250 participants included between April 2010 and November 2011 in 4 different centres in Austria. Inclusion criteria: AMI with or without ST‐segment elevation with imminent immediate invasive intervention and had given written informed consent to participate in the study protocol prior to treatment. Diagnosis of AMI based on clinical symptoms, elevated serum CK or cardiac troponin I levels (or both) and standardised ECG changes. Treating interventional cardiologist had to have expertise for both techniques and had to perform both approaches on a regular basis for diagnostic and intervention purposes (> 150 PCI per year, including at least 40% radial procedures within previous year). Dual circulation of hand assessed by Allen's test in all participants prior to study inclusion. Exclusion criteria: pathological Allen's test, cardiogenic shock with need for implantation of an intra‐aortic balloon pump, any type of vasculitis leading to ischaemia, AV fistula for haemodialysis, peripheral artery disease precluding femoral access, CABG with bilateral mammary artery grafts.
Interventions	Group 1: TRA (n = 125) Group 2: TFA (n = 125)
Outcomes	Primary outcome: occurrence of haematoma, pseudoaneurysm, or local bleeding at access site requiring a subsequent intervention or prolongation of the index hospital stay (or both). Secondary outcomes: 30‐day and 1‐year composite morbidity, defined as overall death or recurrent AMI or stroke procedural times (symptom onset, first medical presentation, arrival at catheter laboratory, insertion of sheath, first injection of contrast agent, balloon inflation and thrombus aspiration), and relevant time intervals were calculated. radiation exposure to participant and operator amount of contrast agent used Clinical follow‐up performed at 30 days and at 1 year.
Notes	Austria Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	High risk	Lost to follow‐up 20%, without clearly accounting for the dropouts: "The 1‐year follow up data were available from 199 patients."
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted

Methods	Parallel‐group RCT. Randomisation to TRA or TFA performed by opening a sealed envelope containing a code for TR or TF angioplasty. Envelopes ordered at random and contained an inclusion number.
Participants	Inclusion criteria: stable and unstable angina pectoris (Braunwald class 1 and 2), type A and B lesions (NHLBI criteria); type C lesions only if there was an intention and technical possibility to implant a stent; multi‐vessel disease; and multi‐vessel PTCA if not more than 1 treated vessel remained unstented. Exclusion criteria: people with an AMI, with unstable angina Braunwald class 3, type C lesion, and chronic total occlusion with anticipated difficult stenting, expected haemodynamic collapse in case of reocclusion, last remaining vessel or unprotected left main PTCA, intracoronary thrombus, any reason for using catheter equipment > 6F (i.e. for non‐balloon technique), non‐PTCA‐related reason for hospitalisation, negative Allen's test for adequate collateral blood supply of the hand, and inability or refusal to give informed consent.
Interventions	Group 1: TRA (n = 322) Group 2: TFA (n = 322)
Outcomes	Primary outcomes: (all within 24 h) cardiac death emergency CABG MI (CK > 2 × upper limit of normal, new Q‐wave) re‐PTCA readmission for unstable angina major entry site complication: radial artery occlusion, AV fistula (false aneurysm), major bleeding (fall in Hb > 2 mmol/L) Secondary outcomes: cost‐effectiveness participant comfort Other outcomes: adverse cardiac event transfusion surgery The above outcomes were reported at 24 h of follow‐up outpatient and inhospital, in addition the study reported 1‐month follow‐up of the following outcomes: death acute MI CABG PTCA
Notes	The Netherlands Funding source: Benelux, Boston Scientific
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of sequence generation
Allocation concealment (selection bias)	Low risk	"Randomization was performed by opening a sealed envelope containing a code for either transradial or transfemoral angioplasty."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low regarding the following outcomes due to being objective and unlikely to be influenced by lack of blinding: cardiac death emergency CABG MI re‐PTCA readmission for unstable angina major entry site complication cost‐effectiveness adverse cardiac event transfusion surgery
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant comfort due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	Unclear risk	Study did not report secondary outcomes prespecified in the methods section (cost‐effectiveness and participant comfort).
Other bias	Low risk	None noted.

Methods	Multi‐centre parallel‐design RCT. Before start of angiography, participants centrally allocated (1:1) to radial or femoral access for diagnostic angiography and PCI, if indicated, using a web‐based system to ensure adequate concealment of allocation. Randomisation sequence was computer generated, blocked and stratified by intended new or ongoing use of ticagrelor or prasugrel, type of ACS (ST‐segment elevation MI, troponin positive or negative, non‐ST‐segment elevation ACS), and anticipated use of immediate PCI. Outcome assessors masked to allocated stent, whereas participants and treating physicians were not.
Participants	8404 people with ACS who were about to undergo CA and possible PCI, if indicated. Inclusion criteria: ACS with or without ST‐segment elevation MI, about to undergo an invasive approach, and interventional cardiologist was willing to proceed with either radial or femoral access and had expertise for both, including ≥ 75 coronary interventions performed, and ≥ 50% of interventions in ACS via the radial route during the previous year. Participants presenting with non‐ST‐segment elevation ACS were eligible if they had a history consistent with new or worsening ischaemia, occurring at rest or with minimal activity within 7 days before randomisation, and fulfilled at least 2 high‐risk criteria. People with ST‐segment elevation MI were eligible if they presented within 12 h of the onset of symptoms or 12‐24 h after onset if there was evidence of continuing ischaemia or previous fibrinolytic treatment, and if they had ST‐segment elevation of ≥ 1 mm in ≥ 2 contiguous leads, new LBBB, or true posterior MI. People with cardiogenic shock, severe peripheral vascular disease, or previous CABG surgery were eligible. Exclusion criteria: use of low molecular weight heparin in previous 6 h, GP IIb/IIIa inhibitors in previous 3 days or any PCI in previous 30 days.
Interventions	Group 1: TRA (n = 4197) Group 2: TFA (n = 4207)
Outcomes	Primary outcomes: MACE: defined as composite of all‐cause mortality, MI, or stroke NACE: defined as composite of major bleeding not related to CABG surgery (BARC type 3 or 5) or MACE. Secondary outcomes: each component of the composite outcomes cardiovascular mortality stent thrombosis: defined as definite or probable occurrence of a stent‐related thrombotic event according to the Academic Research Consortium classification fluoroscopy time and DAP acute kidney injury All outcomes assessed at 30 days.
Notes	Italy, the Netherlands, Spain, and Sweden Funding source: The Medicines Company and Terumo We attempted to contact the authors for raw data to be included in our subgroup analyses, but we could not obtain a valid e‐mail address.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The randomisation sequence was computer generated, blocked and stratified."
Allocation concealment (selection bias)	Low risk	"Before start of angiography, patients were centrally allocated (1:1) to radial or femoral access using a web‐based system to ensure adequate concealment of allocation."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	"Outcome assessors were masked to the allocated stent, whereas patients and treating physicians were not." We judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	"Outcome assessors were masked to the allocated stent." "An independent clinical events committee masked to treatment allocation adjudicated all suspected outcome events by reviewing relevant medical records after site monitoring by Trial Form Support (Lund, Sweden) in Italy and the Netherlands, FLS‐Research Support (Barcelona, Spain) in Spain, and Gothia Forum (Västra Götaland, Sweden) in Sweden."
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Complete follow‐up to 30 days available in 4183 participants in TRA group and 4191 participants in TFA group Dropouts: total 0.35%, TRA 0.33%, TFA 0.38%
Selective reporting (reporting bias)	Low risk	Study protocol available and all study's prespecified (primary and secondary) outcomes of interest in review were reported in prespecified way (ClinicalTrials.gov Identifier: NCT01433627).
Other bias	Low risk	None noted

Methods	Parallel‐group RCT. Eligible participants were 1:1 randomly assigned to TFA group or TRA group by computer generation (in 2 blocks in a 1:1 ratio).
Participants	119 consecutive participants undergoing STEMI who were to receive routine early PCI within 12 h after thrombolysis were enrolled from July 2008 to December 2010. Inclusion criteria: typical clinical presentation; ST elevation of > 0.2 mm in ≥ 2 adjacent precordial leads or 0.1 mm in adjacent limb leads, or new LBBB; received intravenous thrombolysis within 6 h from symptom onset in the non‐PCI hospital; and admitted to hospital within 12 h after intravenous thrombolysis. Exclusion criteria: contradictions of thrombolysis; history of CABG; cardiogenic shock; known difficulties with TFA (i.e. Leriche syndrome, severe peripheral artery disease, large abdominal aortic aneurysm) or TRA (i.e. Raynaud's syndrome); pathological Allen's test; necessity for a preprocedural implantation of a transient pacemaker or intra‐aortic balloon pump; chronic renal insufficiency (creatinine ≥ 2.0 mg/dL) with the potential necessity of using the radial artery as a native fistula in the future; people on haemodialysis with AV fistula; or refusal to participate.
Interventions	Group 1: TRA (n = 60) Group 2: TFA (n = 59)
Outcomes	Puncture time Cannulation time Total procedure time Puncture success rate Time between PCI and ambulation Vascular access‐site complications: minor bleeding (haematoma formation < 5 cm in diameter not requiring specific therapy), major bleeding (Hb loss ≥ 2 g/L, need of blood transfusion, vascular repair, or prolonged hospitalisation), pseudoaneurysm, and artery occlusion Inhospital MACE: defined as death, recurrent MI, and repeat TVR Angiographic characteristics (fluoroscopy time, contrast volume)
Notes	China Funding source: not described
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Eligible patients were 1:1 randomly assigned to transfemoral (TFI group) or transradial catheterisation (TRI group) by a computer generation (in 2 blocks in a 1:1 ratio)."
Allocation concealment (selection bias)	Unclear risk	No description of allocation concealment.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Study did not report blinding, but we judged risk of bias as low since it is an operative procedure and outcomes unlikely to be influenced by lack of blinding of participants and personnel.
Blinding of outcome assessment (detection bias)   Objective outcomes	Low risk	Angiographic results evaluated by 2 independent cardiologists who were blinded to the procedures with the use of qualitative angiographic analysis. Regarding all other outcomes, study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as low due to outcomes being objective and unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias)   Participant‐reported outcomes	High risk	Study did not report blinding; however, blinding impossible due to nature of intervention, but we judged risk of bias as high regarding participant‐reported outcomes due to being subjective and likely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Study reported no loss to follow‐up or withdrawals.
Selective reporting (reporting bias)	Low risk	Study protocol unavailable, but we judged risk of bias as low due to reporting key outcomes expected to be reported for such a study.
Other bias	Low risk	None noted.