Totally percutaneous versus surgical cut‐down femoral artery access for elective bifurcated abdominal endovascular aneurysm repair

Abstract

Background

Abdominal aortic aneurysms (AAAs) are a vascular condition with significant risk attached, particularly if they rupture. Therefore, it is critical to identify and repair these as an elective procedure before they rupture and require emergency surgery. Repair has traditionally been an open surgical technique that required a large incision across the abdomen. Endovascular abdominal aortic aneurysm repairs (EVARs) are now a common alternative. In this procedure, the common femoral artery is exposed via a cut‐down approach and a graft is introduced to the aneurysm in this way. This Cochrane Review examines a totally percutaneous approach to EVAR. This technique gives a minimally invasive approach to femoral artery access that may reduce groin wound complication rates and improve recovery time. However, the technique may be less applicable in people with, for example, groin scarring or arterial calcification. This is an update of the previous Cochrane Review published in 2017.

Objectives

To evaluate the benefits and harms of totally percutaneous access compared to cut‐down femoral artery access in people undergoing elective bifurcated abdominal endovascular aneurysm repair (EVAR).

Search methods

We used standard, extensive Cochrane search methods The latest search was 8 April 2022.

Selection criteria

We included randomised controlled trials in people diagnosed with an AAA comparing totally percutaneous versus surgical cut‐down access endovascular repair. We considered all device types. We only considered studies investigating elective repairs. We excluded studies reporting emergency surgery for ruptured AAAs and those reporting aorto‐uni‐iliac repairs.

Data collection and analysis

We used standard Cochrane methods. Our primary outcomes were 1. short‐term mortality, 2. failure of aneurysm exclusion and 3. wound infection. Secondary outcomes were 4. major complications (30‐day or in‐hospital); 5. medium‐ to long‐term (6 and 12 months) complications and mortality; 6. bleeding complications and haematoma; and 7. operating time, duration of intensive treatment unit (ITU) stay and hospital stay. We used GRADE to assess the certainty of evidence for the seven most clinically relevant primary and secondary outcomes.

Main results

Three studies with 318 participants met the inclusion criteria, 189 undergoing the percutaneous technique and 129 treated by cut‐down femoral artery access. One study had a small sample size and did not adequately report the method of randomisation, allocation concealment or preselected outcomes. The other two larger studies had few sources of bias and good methodology; although one study had a high risk of bias in selective reporting.

We observed no clear difference in short‐term mortality between groups, with only one death occurring overall, in the totally percutaneous group (risk ratio (RR) 1.50, 95% confidence interval (CI) 0.06 to 36.18; 2 studies, 181 participants; low‐certainty evidence). One study reported failure of aneurysm exclusion. There was one failure of aneurysm exclusion in the surgical cut‐down femoral artery access group (RR 0.17, 95% CI 0.01 to 4.02; 1 study, 151 participants; moderate‐certainty evidence). For wound infection, there was no clear difference between groups (RR 0.18, 95% CI 0.01 to 3.59; 3 studies, 318 participants; moderate‐certainty evidence).

There was no clear difference between percutaneous and cut‐down femoral artery access groups in major complications (RR 1.21, 95% CI 0.61 to 2.41; 3 studies, 318 participants; moderate‐certainty evidence), bleeding complications (RR 1.02, 95% CI 0.29 to 3.64; 2 studies, 181 participants; moderate‐certainty evidence) or haematoma (RR 0.88, 95% CI 0.13 to 6.05; 2 studies, 288 participants).

One study reported medium‐ to long‐term complications at six months, with no clear differences between the percutaneous and cut‐down femoral artery access groups (RR 0.82, 95% CI 0.25 to 2.65; 1 study, 135 participants; moderate‐certainty evidence).

We detected differences in operating time, with the percutaneous approach being faster than cut‐down femoral artery access (mean difference (MD) −21.13 minutes, 95% CI −41.74 to −0.53 minutes; 3 studies, 318 participants; low‐certainty evidence). One study reported the duration of ITU stay and hospital stay, with no clear difference between groups.

Authors' conclusions

Skin puncture may make little to no difference to short‐term mortality. There is probably little or no difference in failure of aneurysm exclusion (failure to seal the aneurysms), wound infection, major complications within 30 days or while in hospital, medium‐ to long‐term (six months) complications and bleeding complications between the two groups. Compared with exposing the femoral artery, skin puncture may reduce the operating time slightly. We downgraded the certainty of the evidence to moderate and low as a result of imprecision due to the small number of participants, low event rates and wide CIs, and inconsistency due to clinical heterogeneity. As the number of included studies was limited, further research into this technique would be beneficial.

Plain language summary

Is skin puncture better than exposing the femoral artery for minimally invasive repairs of abdominal aortic aneurysms?

Key messages

– We found insufficient high‐quality evidence about whether skin puncture or exposing the femoral artery is better for minimally invasive repairs of abdominal aortic aneurysms.

– Skin puncture may make little to no difference to short‐term mortality. There is probably little or no difference in failure to seal the aneurysms, wound infection, major complications within 30 days or while in hospital, medium‐ to long‐term (six months) complications and bleeding complications between the two groups. Compared with exposing the femoral artery, skin puncture may reduce the operating time slightly.

– More, larger, well‐designed studies are needed to give better estimates of the benefits and potential harms of two different access methods for minimally invasive repairs of abdominal aortic aneurysms.

What is an abdominal aortic aneurysm?

An abdominal aortic aneurysm is a ballooning of the largest blood vessel in the abdomen, the abdominal aorta, due to weakness of the vessel wall. This ballooning may lead to life‐threatening rupture. Repair of the aneurysm is recommended if the risk of rupture is greater than the risk of surgery.

How is an abdominal aortic aneurysm treated?

Most repairs involve putting in an artificial graft, a tube composed of fabric, to help strengthen the artery wall. There are two main methods for repair. One is an open technique in which the whole abdomen is opened and the graft is used to replace the diseased part of the vessel. The other technique is endovascular aneurysm repair. With this minimally invasive technique, the graft is fed into the abdominal aorta through an artery in the groin (the femoral artery) avoiding the large abdominal incision. This review looked at an alternative method for introducing the graft into the femoral artery, percutaneous access. Instead of making an incision in the groin to expose the femoral artery (a cut‐down), a needle is inserted into the femoral artery and a flexible guide wire is inserted through the needle. The needle is removed and a plastic tube is introduced into the femoral artery over the guide wire (percutaneous access), with a small cut in the skin to allow the passage of the plastic tube. Once introduced, the guide wire can be removed leaving the tube in place in the artery. The artificial graft and all other materials can then be fed into the artery via the plastic tube. Once the procedure is complete the tube can be withdrawn. The surface incision can usually be closed with a single stitch.

What did we want to find out?

We wanted to know if skin puncture was better than exposing the femoral artery for minimally invasive repairs of abdominal aortic aneurysms.

What did we do?

We searched for studies that included people with abdominal aortic aneurysms who were treated with minimally invasive repairs. These studies randomly selected participants to receive the minimally invasive repairs by skin puncture or exposing the femoral artery. We compared and summarised their results and rated our confidence in the evidence, based on factors such as study methods and sizes.

What did we find?

We found three studies involving 318 people with abdominal aortic aneurysms who were treated with minimally invasive repairs by skin puncture or exposing the femoral artery. We have little confidence that there is no difference in short‐term mortality and are moderately confident that there is no difference in failure to seal the aneurysms, wound infection, major complications within 30 days or while in hospital, medium‐ to long‐term (six months) complications and bleeding complications. We have little confidence that skin puncture reduces the operating time slightly.

What are the limitations of the evidence?

We have moderate to little confidence in the evidence because the results were imprecise due to the small number of participants and low event rate, and the studies used different types of devices for delivering artificial grafts.

How up to date is this evidence?

This review updates a previous Cochrane Review. The evidence is up to date to April 2022.

Summary of findings

Summary of findings 1. Totally percutaneous compared to cut‐down femoral artery access for elective bifurcated abdominal endovascular aneurysm repair.

Totally percutaneous compared to cut‐down femoral artery access for elective bifurcated abdominal endovascular aneurysm repair
Patient or population: people undergoing elective bifurcated abdominal endovascular aneurysm repair Setting: hospital Intervention: totally percutaneous Comparison: cut‐down femoral artery access
Outcomes	Anticipated absolute effects* (95% CI)			Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with cut‐down femoral artery access	Risk with totally percutaneous	Difference (95% CI)
Short‐term mortality (30‐day or in‐hospital)	See comment	See comment	See comment	RR 1.50 (0.06 to 36.18)	181 (2 RCTs)	⊕⊕⊝⊝ Lowa	It was not possible to calculate risk as there was only 1 event. Note that although 2 RCTs were included, only 1 contributed to the effect estimate (no events in Torsello 2003).
Failure of aneurysm exclusion (follow‐up 28 days)	Study population			RR 0.17 (0.01 to 4.02)	151 (1 RCT)	⊕⊕⊕⊝ Moderateb	—
	20 per 1000	3 per 1000 (0 to 80)	−17 (−20 to 60)
Wound infection (30‐day or in‐hospital)	Study population			RR 0.18 (0.01 to 3.59)	318 (3 RCTs)	⊕⊕⊕⊝ Moderatec	Note that although 3 RCTs were included, only 1 contributed to the effect estimate (0 events in Torsello 2003 and Nelson 2014).
	16 per 1000	3 per 1000 (0 to 56) See comment	−13 (−16 to 40)
Major complications (30‐day or in‐hospital)	Study population			RR 1.21 (0.61 to 2.41)	318 (3 RCTs)	⊕⊕⊕⊝ Moderated	—
	140 per 1000	169 per 1000 (85 to 336)	29 (−55 to 196)
Medium‐ to long‐term complications (follow‐up 6 months)	Study population			RR 0.82 (0.25 to 2.65)	135 (1 RCT)	⊕⊕⊕⊝ Moderatee	—
	93 per 1000	76 per 1000 (23 to 247)	−17 (−70 to 154)
Bleeding complications (30‐day or in‐hospital)	Study population			RR 1.02 (0.29 to 3.64)	181 (2 RCTs)	⊕⊕⊕⊝ Moderatef	—
	46 per 1000	47 per 1000 (13 to 168)	1 (−33 to 122)
Operating time (minutes)	The mean operating time was 129.2 minutes	The mean operating time in the intervention group was 21.13 minutes fewer (41.74 fewer to 0.53 fewer) See comment	−21.13 (−41.74 to −0.53)	—	318 (3 RCTs)	⊕⊕⊝⊝ Lowg	The CIs included trivial effect (0.53 minutes fewer) and important effect (41.74 minutes fewer).
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: 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 certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded two levels for very serious imprecision (low number of participants and events, extremely wide confidence intervals included both harm and benefit).
^bDowngraded one level for serious imprecision (low number of participants and events, wide confidence intervals included both harm and benefit, and data were available from only one study).
^cDowngraded one level for serious imprecision (low number of participants and events, wide confidence intervals included both harm and benefit).
^dDowngraded one level for serious imprecision (low number of participants and events, wide confidence intervals included both harm and benefit).
^eDowngraded one level for serious imprecision (low number of participants and events, wide confidence intervals included both harm and benefit, and data were available from only one study). There was a loss to follow‐up for this outcome (10% of participants; 8% of the percutaneous group and 14% of the cut‐down femoral artery access group). We did not downgrade further as there was a clear breakdown of loss to follow‐up and no clear difference in the loss to follow‐up between groups was detected.
^fDowngraded one level for serious imprecision (low number of participants and events, wide confidence intervals included both harm and benefit).
^gDowngraded one level for serious imprecision (the studies were not adequately powered for this outcome) and one level for serious inconsistency (I² = 73% due to clinical heterogeneity).

Background

Description of the condition

Abdominal aortic aneurysms (AAAs) are a vascular condition whereby the abdominal aorta increases in diameter. This enlargement is based on the abdominal aorta's diameter, which is 50% greater than the normal aortic diameter (approximately 2 cm), or larger than 3 cm. Significant risk comes with the enlargement: if untreated, there is a high risk of rupture, which carries an approximately 80% risk of mortality (Chambers 2009). Therefore, it is critical to identify and repair these aneurysms as an elective procedure before they rupture and require emergency surgery.

In some European countries such as Sweden (Svensjö 2011) and the UK (Owens 2019), the prevalence of AAAs among screened men aged 65 years or older has been estimated to range from 1.2% to 2.7% (Guirguis‐Blake 2019). Recent systematic reviews have shown the prevalence of AAAs ranged from 1.01% to 1.59% in Asian people (Chan 2021), and 0.7% to 6.4% in African people (Ngetich 2020). Meanwhile, studies have consistently demonstrated that females have a far lower prevalence (Kent 2010; Vardulaki 2000). Additionally, advancing age, smoking and a family history of AAAs are the main risk factors for AAAs (Guirguis‐Blake 2019; Jacomelli 2017; Jahangir 2015; Kent 2010; Summers 2021; Ullery 2018).

Description of the intervention

Repair has traditionally been by open surgery requiring a large incision across the abdomen to allow clear access to the aorta so that the damaged section could be repaired by the insertion of a prosthetic graft. In 1991, Parodi and colleagues reported their initial experiences with what they described as a transfemoral intraluminal graft (Parodi 1991). This rapidly developed into what we now refer to as endovascular abdominal aortic aneurysm repair (EVAR). Currently, EVARs and open repairs are the two primary forms of treatment for AAAs.

In this procedure, the common femoral arteries are exposed via incisions (i.e. the direct surgical femoral artery cut‐down) in each groin, and guidewires and access sheaths are introduced. Then, under radiographic control, the components of the endovascular graft are advanced over the guidewires and deployed. The use of angiography is performed to confirm satisfactory deployment of the prosthesis and exclusion of the aneurysm sac, following which all catheters, sheaths and guidewires are removed. The defects in the femoral arteries are surgically repaired and the cut‐down incisions closed (Rutherford 2005).

EVAR has grown rapidly in popularity and in the US, between 2003 and 2006, accounted for the majority of procedures taking place (Schwarze 2009). In the UK, the National Vascular Registry reported that EVAR is currently used in the majority (61.5%) of elective infra‐renal AAA repairs with 5851 elective EVAR repairs conducted between 2018 and 2020 (Waton 2022). It was also used in 37.7% of ruptured AAA repairs between 2019 and 2020 (Waton 2022).

More recently, further advances have taken place allowing for a totally percutaneous approach. This approach was reported as early as 1999 (Papazoglou 1999) and has grown in popularity with an analysis of NSQIP (National Surgical Quality Improvement Program) data showing almost half of elective EVAR cases reviewed used a percutaneous approach (Kauvar 2016). The approach has been developed using several different suture‐mediated closure devices being reported in the literature including the Prostar XL (PS; Abbott Vascular, Redwood City, California, US) (Malkawi 2010) and ProGlide system (PG; Abbott Vascular) for closure (Dosluoglu 2007). The adoption may also have been influenced by the rise of low‐profile stent‐grafts, which are more malleable, easier to introduce and navigate, and have been shown to compare favourably in the mid‐term in people with challenging anatomy (Sobocinski 2015).

The most common technique employed is the preclose technique. Access is gained to the common femoral artery (CFA) and the initial sheath is replaced by a percutaneous closure device over a guidewire. The major difference in this technique from percutaneous techniques used in other procedures is the deployment of a suture‐mediated closure device at the start of the procedure. This allows the arteriotomy to be enlarged, enabling the insertion of sheaths of a larger size and allowing the EVAR to be conducted normally. At the conclusion of the repair, the sheath is then removed using a knot pusher to aid closure. The final wound is closed using a single suture or tape. Some variations of this technique exist at different centres and using different devices (Lee 2008; Malkawi 2010; Watelet 2006).

How the intervention might work

The stent grafts used and the manner in which they achieve aneurysm exclusion are identical to the already widely used EVAR techniques. As such, many of the long‐term complications are similar.

The key differences are the initial arterial access, wound closure and the use of a percutaneous suture‐mediated closure device during surgery. One concern with EVAR procedures is the continued problem of wound closure, which a number of studies have examined (Dalainas 2004; Faries 2002). As the percutaneous approach is less invasive and results in smaller final wound closure, it may help to address this problem. Wound infection rates with percutaneous closure devices can be as low as 0.2% (Watelet 2006), compared with 2% in cut‐down femoral artery access (Slappy 2003).

Some people may be at risk of problems with arterial access using this intervention. These include people with calcification of the femoral artery, groin scarring or an inguinal arterial prosthesis, morbid obesity and small or tortuous arterial iliac arteries (Traul 2000; Watelet 2006). This makes participant selection key to the initial success of the procedure. In addition, the percutaneous technique may not always achieve adequate haemostasis (Torsello 2003), and there may be an increased risk of bleeding and haematoma formation with a need to convert to a conventional cut‐down approach.

Why it is important to do this review

It has been suggested that a totally percutaneous approach to EVAR offers a less invasive treatment with faster recovery and reduced wound infection rates than with cut‐down femoral artery access. This hypothesis must be tested by a robust systematic review as this would make percutaneous access a preferable choice for many surgical interventions. In the review, we aimed to evaluate the efficacy and safety of percutaneous access as an alternative procedure and highlight any shortcomings in the technique. This review is important to allow clinicians to make a more informed decision when deciding on the appropriate technique for each individual patient.

This review was originally conducted in 2014 (Jackson 2014), with one included and one ongoing trial identified that met the inclusion criteria; and subsequently updated in 2017 with two eligible included and two eligible ongoing trials (Gimzewska 2017). It is good practice to formally update the review to include any newly available data on a regular basis.

Objectives

To evaluate the benefits and harms of totally percutaneous access compared to cut‐down femoral artery access in people undergoing elective bifurcated abdominal endovascular aneurysm repair (EVAR).

Methods

Criteria for considering studies for this review

Types of studies

We considered only randomised controlled trials (RCTs) for evaluation. We would have included cross‐over trials, with only the first phase results considered, but did not identify any.

Types of participants

We included people diagnosed with an AAA that had been visualised using either computed tomography (CT) or ultrasonography.

We considered only people undergoing elective repairs. We excluded people undergoing emergency surgery for a ruptured AAA and those undergoing aorto‐uni‐iliac (also known as aorto‐mono‐iliac) repairs.

Types of interventions

The primary intervention was a totally percutaneous endovascular repair. We considered all device types. We compared this against cut‐down femoral artery access for endovascular repair.

Types of outcome measures

Primary outcomes

• Short‐term mortality (30‐day or in‐hospital, i.e. procedure‐related)

• Failure of aneurysm exclusion (defined as no flow into the AAA sac on follow‐up imaging) 30 days after the procedure

• Wound infection (30‐day or in‐hospital)

With regard to the primary outcomes of mortality and failure of aneurysm exclusion, at least equivalence but not necessarily superiority over surgical cut‐down had to be demonstrated to allow a recommendation as percutaneous interventions may still be preferable if there are fewer wound problems and length of hospital stay is reduced.

Secondary outcomes

• Major complications, for example, myocardial infarction, stroke, renal failure, respiratory failure, lower limb ischaemia, pneumonia, technical failure or conversion to alternative technique (30‐day or in‐hospital)

• Medium‐ to long‐term (6 and 12 months) complications and mortality (re‐interventions, device‐related complications, cause of death)

• Bleeding complications and haematoma (30‐day or in‐hospital)

• Operating time (minutes), duration of intensive treatment unit (ITU) stay (hours) and hospital stay (days)

Search methods for identification of studies

Electronic searches

The Cochrane Vascular Information Specialist conducted systematic searches of the following databases for randomised controlled trials and controlled clinical trials without language, publication year or publication status restrictions:

• Cochrane Vascular Specialised Register via the Cochrane Register of Studies (CRS‐Web);

• Cochrane Central Register of Controlled Trials (CENTRAL; 2022) via the Cochrane Register of Studies Online (CRSO);

• MEDLINE (Ovid MEDLINE Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE Daily and Ovid MEDLINE);

• Embase Ovid;

• CINAHL EBSCO.

We developed search strategies for other databases from the search strategy designed for MEDLINE. Where appropriate, they were combined with adaptations of the highly sensitive search strategy designed by Cochrane for identifying randomised controlled trials and controlled clinical trials (as described in the Cochrane Handbook for Systematic Reviews of Interventions Chapter 4, Lefebvre 2022). Search strategies for major databases are provided in Appendix 1.

We searched the following trials registries:

• World Health Organization International Clinical Trials Registry Platform (www.who.int/clinical-trials-registry-platform);

• ClinicalTrials.gov (clinicaltrials.gov).

The most recent searches were carried out on 8 April 2022.

Searching other resources

We checked reference lists of retrieved articles for any additional eligible studies.

Data collection and analysis

We conducted data collection and analysis in accordance with the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022).

Selection of studies

Pairs of review authors (from QW, YM, YZ, XS, JW, FL) independently screened all titles and abstracts identified from the literature searches and excluded irrelevant studies using Covidence (Covidence). We retrieved those studies deemed relevant in full texts. Pairs of review authors (from QW, JW, YM, YZ) independently assessed the full‐text articles for suitability for inclusion. We resolved all disagreements by consensus and, when necessary, we consulted another two review authors (FL, LY). We contacted trial registries to request further details for one record with missing information. For non‐English language publications, we used Google Translate (translate.google.com) and DeepL Translator (www.deepl.com/translator) for the initial eligibility assessment, then requested confirmation with help from native speakers.

We presented screening and selection processes in a PRISMA flow diagram (Page 2021a; Page 2021b), and listed all articles excluded at the full‐text stage, with their exclusion reasons, in the Characteristics of excluded studies table.

Data extraction and management

Pairs of review authors (from QW, YM, YZ, XS) independently extracted data from the included studies using a modified data extraction form provided by Cochrane Vascular. To ensure the understanding of extraction items among all review authors, we conducted a pilot data extraction for one study. We resolved all disagreements through discussions and, when necessary, we consulted another three review authors (JW, FL, LY). If necessary, we would have contacted the study authors for clarification.

Assessment of risk of bias in included studies

Pairs of review authors (from QW, XS, SX, ML) independently assessed included studies for risk of bias using Cochrane's RoB 1 tool (Higgins 2017). This tool covers sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting and other sources of bias. We assessed each of these domains as high, low or unclear risk of bias and provided support for each judgement. We resolved all disagreements through discussion and, when necessary, we consulted another two review authors (FL, LY). We presented our conclusions in the risk of bias figures and risk of bias table.

Measures of treatment effect

We measured the treatment effect according to the statistical guidelines recommended by Cochrane Vascular. For dichotomous data, we calculated risk ratios (RRs) with 95% confidence intervals (CIs) as the measure of effect. We analysed continuous data using mean differences (MD). If studies reported similar outcomes on different scales, we planned to calculate standardised mean difference (SMD).

Unit of analysis issues

The primary analysis was per participant randomised. Our search did not identify any cross‐over trials but if it had, we would only have included data from the first phase only.

Dealing with missing data

There were no missing data relating to primary outcomes. For each included study, we reported the details (including the rate of loss to follow‐up and reasons) if participants were excluded after allocation. Had data been missing we would have contacted the study authors to request them. If this was not possible, we planned to use available‐case data in the analyses without imputing values for missing data.

Assessment of heterogeneity

We quantified statistical heterogeneity using the I² statistic (Higgins 2003) with the following interpretations based on the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022): 0% to 40%: might not be important; 30% to 60%: may represent moderate heterogeneity; 50% to 90%: may represent substantial heterogeneity; 75% to 100%: considerable heterogeneity.

When observing substantial heterogeneity with I² > 50%, we planned to explore the reasons by examining the characteristics of included articles, including participants, interventions and study design (Deeks 2022).

Assessment of reporting biases

The single best step undertaken to minimise reporting bias was a full and comprehensive search encompassing both published and grey literature, whilst being aware of duplication of data. When including more than 10 studies within the final analysis, we would have used a funnel plot to explore small‐study effects (Egger 1997; Page 2021b). Also, we examined the protocols and registration information for included trials against published reports to explore the potential selective reporting.

Data synthesis

We synthesised available data using Review Manager Web software (RevMan Web 2022). Two review authors (QW, JW) performed data analyses based on the instructions in Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022). Given the potential moderate and substantial heterogeneity that might be identified, we used a random‐effects model to investigate pooled estimates of the effects of treatment by calculating RRs with 95% CIs for dichotomous outcomes and MDs or SMDs with 95% CIs for continuous outcomes.

Had we identified considerable heterogeneity among the included studies (I² = 75% to 100%), we would have used a narrative approach for data synthesis rather than pooling results in a meta‐analysis.

Subgroup analysis and investigation of heterogeneity

We did not plan or perform any subgroup analyses owing to the small number of studies included. If there had been substantial statistical heterogeneity identified (I² > 50%), we would have explored the potential clinical reasons.

Sensitivity analysis

We intended to carry out sensitivity analyses to explore the robustness of the conclusions reached. These would have examined several factors, including methodological quality and risk of bias, data source (published or unpublished), and any other analyses that might have appeared significant upon completion of the review. Given the low number of trials (three) and low risk of bias, a sensitivity analysis was not deemed appropriate.

Summary of findings and assessment of the certainty of the evidence

We created a summary of findings table to present the main results of this review using GRADEpro GDT software (GRADEpro GDT; Schünemann 2022a). We used the GRADE approach to assess the certainty of evidence for the seven primary and secondary outcomes considered to be most clinically relevant to healthcare professionals and patients (Guyatt 2008), as follows.

• Short‐term mortality (30‐day or in‐hospital)

• Failure of aneurysm exclusion (30 days after the procedure)

• Wound infection (30‐day or in‐hospital)

• Major complications (30‐day or in‐hospital)

• Medium‐ to long‐term (6 and 12 months) complications and mortality

• Bleeding complications (30‐day or in‐hospital)

• Operating time (minutes)

The body of evidence from randomised trials began with a high‐certainty rating and could be downgraded to moderate, low or very low based on the risk of bias, indirectness of evidence, inconsistency of results, imprecision of effect estimates and potential publication bias according to the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2022b).

Results

Description of studies

See Characteristics of included studies and Characteristics of excluded studies tables.

Results of the search

See Figure 1.

1.

Flow diagram

The searches at this update identified 2003 articles. After deduplication, we screened 1462 articles in Covidence (Covidence). We assessed 20 full‐text articles for eligibility. We identified one new study (Vierhout 2019), with two additional publications (Vierhout 2015; Vierhout 2016) and two additional publications to the previously included study Nelson 2014 (NCT01070069; Nelson 2012). We identified six new excluded studies (Iłżecki 2018; Krajcer 2011; NTR4225; Pitton 2003; Roche‐Nagle 2018; Uhlmann 2018) with two additional publications for Uhlmann 2018 (NCT02822560; Uhlmann 2017). We did not find any new ongoing studies.

Included studies

We included one new study (three reports) in this update (Vierhout 2019). Therefore, three independent studies met the inclusion criteria (Nelson 2014; Torsello 2003; Vierhout 2019).

Nelson 2014 presented the PEVAR (totally percutaneous access versus open femoral exposure for endovascular aortic aneurysm repair) trial. This was a multicentre randomised controlled trial conducted in 18 centres in the US. All centres had prior experience in EVAR using both cut‐down femoral artery access and percutaneous techniques. All investigators (surgeons and interventionalists) were required to provide evidence of experience in the use of closure devices, complete the Endologix EVAR training programme and perform 20 PEVAR preclose cases before commencing the trial.

The trial enrolled 151 participants (136 men, 15 women). Anyone over the age of 18 years presenting with an AAA with a maximal diameter of 5 cm or more, or rapidly expanding was considered. Participants had to understand and sign an informed consent form, and agree to all follow‐up visits. Anatomical eligibility was determined by CT of the thoracic and abdominal aorta. An experienced, independent vascular surgeon conducted additional assessments of CFA characteristics. Participants were only included if they had an absence of the following: prior groin incision, haematoma or significant scarring at the ipsilateral arterial access site, clip‐based vascular closure device placement ever, collagen‐based vascular closure device placement in either arterial access site within 90 days, femoral artery needle puncture in either arterial access site within 30 days, active groin infection, traumatic vascular injury, femoral artery aneurysm or pseudoaneurysm, or arteriovenous fistula. Participants had to have a life expectancy of more than one year, serum creatinine 1.7 mg/dL or less, body mass index (BMI) less than 40 kg/m², have no planned major interventions or surgery within 30 days after the study procedure, and be free from myocardial infarction and cerebrovascular accident within three months of enrolment.

Of 179 screened participants, 153 were suitable, and 151 were enrolled in the trial. Participants were randomised using a 2:1 design to percutaneous access/closure (101 participants) or cut‐down femoral artery access (termed open femoral exposure (FE)) (50 participants). The percutaneous access/closure procedure group was divided equally between two suture‐mediated preclosure device types: the PG (50 participants) and the PS (51 participants) (both Abbott Vascular). All endovascular AAA repairs were performed using the Endologix 21F profile sheath‐based system (Endologix, Inc, Irvine, California, US). Postdeployment graft position/aneurysm exclusion was confirmed using angiography.

The primary endpoint was 'treatment success', a composite endpoint comprising procedural technical success and absence of major adverse events and vascular complications at 30 days. Follow‐up was six‐months.

Torsello 2003 was a small, randomised pilot study conducted in Germany. It consisted of 30 participants (29 men, one woman) aged 51 to 90 years (mean 72.9, standard deviation (SD) 9.9 years). Anyone presenting with an AAA was considered, including those with calcification of the femoral artery, scars in the groin or obesity. The exclusion criteria included people with psychiatric conditions, those undergoing the implantation of an aorto‐uni‐iliac graft and people with an aneurysm of the femoral artery. Aneurysms were repaired using either the Zenith graft (Cook, Bloomington, Indiana, US) (16 participants) or the Talent endovascular graft (Medtronic, Sunrise, Florida, US) (14 participants), although the paper did not describe how these different grafts were distributed among the treatment arms.

In 15 participants undergoing the percutaneous procedure, a PS percutaneous vascular surgery device (Abbott Vascular) was used in the 'preclose' technique for closure of the access site. For 15 participants undergoing cut‐down femoral artery access, a transverse groin incision was made to expose the CFA for direct needle puncture. All participants received heparin 5000 IU after sheath insertion. Duplex ultrasound scanning was performed before and after the procedure.

The paper did not explicitly state which outcomes had been selected but reported a wide range of outcomes, including time of surgery, mortality and major complications.

Vierhout 2019 presented the PiERO (Percutaneous access in Endovascular Repair vs Open) trial. This was a multicentre randomised single‐blind controlled trial conducted at six hospitals in the Netherlands. The trial consisted of 137 participants (123 men, 14 women) aged 56 to 93 years (mean 72 years). All participating vascular surgeons and interventional radiologists had completed at least 20 open CFA access and percutaneous procedures. Furthermore, before participants entered the study, the company that supplied the closure device (Abbott Vascular) reviewed and certified the percutaneous access technique.

Participants aged over 18 years who were scheduled for an elective EVAR with the following eligibility criteria were considered: physically and mentally capable of giving consent, presenting with an AAA exceeding 5.5 cm in diameter, and growth of 5 mm or more within six months. The exclusion criteria included: CFA access sites with over 50% circumferential calcification (based on multidisciplinary discussion and radiologists' report); previous CFA surgery; documented infection at the time of operation; and fewer or more than two accesses (additional brachial or carotid), fenestrated or branched EVAR.

Each participant with an AAA suitable for EVAR was operated through arterial access in both groins and randomised to open or percutaneous access to the main device through the CFA. The access on the contralateral side was operated using the opposite technique. The type of endovascular stent graft was not reported. For 73 participants in the percutaneous group, access to the main device was obtained via ultrasound‐guided puncture of the CFA, followed by positioning of one or more PS or PG devices (Abbott Vascular) through an incision of approximately 1 cm. For 64 participants in the open group, access to the main device was obtained via a 3‐ to 4‐cm craniocaudal incision and puncture of the CFA under direct vision. After completion of angiography and removal of the sheaths, haemostasis was obtained in both groups. All participants in both groups received prophylactic antibiotics 15 to 60 minutes before surgery and intravenous heparin 5000 units during the procedure.

The primary endpoint was the number of surgical site infections (SSIs). The secondary endpoints were wound complications, standardised wound assessment scores/Southampton Wound Assessment score, visual analogue scale (VAS) score, access‐related vascular injuries (ARVI), serious adverse events, duration of surgery and blood loss.

Excluded studies

Six new studies were excluded in this update (Iłżecki 2018; Krajcer 2011; NTR4225; Pitton 2003; Roche‐Nagle 2018; Uhlmann 2018). In total, we excluded 11 studies. One trial did not include the primary intervention (Hattab 2012). Eight trials were non‐randomised (Ichihashi 2016; Iłżecki 2018; Jean‐Baptiste 2008; Krajcer 2010; Krajcer 2011; Pitton 2003; Roche‐Nagle 2018; Xiong 2012). One trial (Uhlmann 2018) and two additional publications (NCT02822560; Uhlmann 2017) included the wrong population. Despite contacting trial registries, we were unable to locate one record for further details (NTR4225). See Characteristics of excluded studies table.

Ongoing studies

We identified no ongoing studies. One study previously assessed as ongoing is now included (Vierhout 2015) and another is excluded (NCT02822560).

Risk of bias in included studies

See Figure 2 and Figure 3.

2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Allocation

Vierhout 2019 reported randomisation prior to the operation using envelopes, which were produced outside the treatment centres, and consisted of 20 envelopes per batch. The participants were randomised to the percutaneous or the open CFA access group in a 1:1 ratio. The allocation was concealed to participants, follow‐up physicians, data collectors and data analysts. However, the operating team (physician or assistant) could not be blinded for the allocation and, therefore, selected an envelope to allocate the participants to one of the two groups. Therefore, Vierhout 2019 was at low risk of selection bias. Nelson 2014 reported randomisation by study site, using two block sizes (three or six) with random choice of block size order. The study used 2:1 randomisation (PEVAR:FE), with equal allocation to the two PEVAR groups (PG, PS). Nelson 2014 provided centres with a set of sealed randomised envelopes. Upon a participant meeting eligibility screening criteria, the next sequential envelope was opened and the assignment immediately allocated. Therefore, Nelson 2014 was at low risk of selection bias. Torsello 2003 did not report either random sequence allocation or allocation concealment. This made it impossible to make a formal judgement on the risk of bias generated and was at unclear risk of selection bias.

Blinding

Vierhout 2019 reported that the operations were concealed to participants and the performing surgeon could not be blinded when deciding the access side for the main device based on the participant's anatomy. The operations were also concealed to follow‐up physicians, data collectors and data analysts. The primary endpoint measurement was objective in nature and was assessed by an independent researcher who was blinded to the main device access site, and the data were analysed anonymously. Neither Nelson 2014 nor Torsello 2003 reported the blinding of trial participants and outcome assessments; it is impossible to blind personnel performing the procedure. We judged all three studies at unclear risk of performance bias. This is unlikely to affect the risk of bias as the outcomes selected by the study authors, and those analysed in this Cochrane Review, were not susceptible to bias from a lack of blinding, so the three included studies were at low risk of detection bias.

Incomplete outcome data

Vierhout 2019 had no loss to follow‐up at six weeks postoperatively with all assigned participants fully reported (low risk of attrition bias). Both Nelson 2014 and Torsello 2003 had complete follow‐up of all participants at 30 days or in‐hospital, and were at low risk of attrition bias.

Nelson 2014 had a 10% loss to follow‐up at six months, which introduced attrition bias for the secondary outcome medium‐ to long‐term complications. Loss to follow‐up was higher in the cut‐down femoral artery access group (7/50, 14%) than in the totally percutaneous group (8/100, 8%); however, we found no clear difference (P = 0.53). Reasons for loss to follow‐up were given for each group. In the totally percutaneous group, six participants withdrew, and two participants were missed. In the cut‐down femoral artery access group, three participants withdrew, two participants were missed and two refused six‐month follow‐up. The same proportion of participants in each group withdrew (6%). Twice as many participants were missed in the cut‐down femoral artery access group (2/50, 4%) compared with the totally percutaneous group (2/100, 2%) and a further two participants (4%) of the cut‐down femoral artery access group refused six‐month follow‐up.

Selective reporting

Vierhout 2019 did not report the number of SSI (the primary endpoint) at one year, which should be evaluated based on the protocol. Also, we found some inconsistencies in results between study reports (Vierhout 2016; Vierhout 2019). Therefore, Vierhout 2019 was at high risk of selective reporting bias. Nelson 2014 clearly stated its primary and secondary outcomes and reported on these fully with raw data presented, and was at low risk of selective reporting bias. Torsello 2003 did not outline its selected outcomes and, therefore, it was prone to a higher risk of bias due to selective reporting of the results. However, major outcomes identified by this review appeared to have been reported so it was at unclear risk of reporting bias.

Other potential sources of bias

There were no other sources of bias identified (low risk of other bias) (Nelson 2014; Torsello 2003; Vierhout 2019).

Effects of interventions

See: Table 1

See Table 1.

Primary outcomes

Short‐term mortality (30‐day or in‐hospital, i.e. procedure‐related)

Two studies reported data on short‐term mortality. Torsello 2003 recorded no short‐term or operative mortality in either group. Nelson 2014 reported one cardiac death on day 28 in the percutaneous group (RR 1.50, 95% CI 0.06 to 36.18; P = 0.8; 2 studies, 181 participants; low‐certainty evidence; Analysis 1.1). Note that although Analysis 1.1 considered two studies, Torsello 2003 (30 participants) reported no mortality and, therefore, did not contribute to the effect estimate in this analysis.

1.1. Analysis.

Comparison 1: Percutaneous versus cut‐down femoral artery access, Outcome 1: Short‐term mortality (30‐day or in‐hospital)

Failure of aneurysm exclusion 30 days after the procedure

One study reported data on failure of aneurysm exclusion. Nelson 2014 conducted follow‐up CT scans at 28 days, with one participant in the cut‐down femoral artery access group requiring a type 1a endoleak repair within one month after the procedure (RR 0.17, 95% CI 0.01 to 4.02; P = 0.27; 151 participants; moderate‐certainty evidence; Analysis 1.2).

1.2. Analysis.

Comparison 1: Percutaneous versus cut‐down femoral artery access, Outcome 2: Failure of aneurysm exclusion

Wound infection (30‐day or in‐hospital)

Three studies reported wound infection. Both Torsello 2003 and Nelson 2014 reported that there were no wound infections in either group. Vierhout 2019 reported no SSI in the percutaneous group and two SSIs in the open access group (P = 0.34) (RR 0.18, 95% CI 0.01 to 3.59; P = 0.26; 3 studies, 318 participants; moderate‐certainty evidence; Analysis 1.3). Note that although Analysis 1.3 considered three studies, Torsello 2003 and Nelson 2014 reported no wound infections and, therefore, they did not contribute to the effect estimate in this analysis.

1.3. Analysis.

Comparison 1: Percutaneous versus cut‐down femoral artery access, Outcome 3: Wound infection (30‐day or in‐hospital)

Secondary outcomes

Major complications (30‐day or in‐hospital)

Three studies reported major complications. When the numbers of participants experiencing adverse events were combined, there was no clear difference in major complications between the percutaneous and cut‐down femoral artery access groups (RR 1.21, 95% CI 0.61 to 2.41; P = 0.59; 3 studies, 318 participants; moderate‐certainty evidence; Analysis 1.4). There was little heterogeneity between the studies (I² = 31%).

1.4. Analysis.

Comparison 1: Percutaneous versus cut‐down femoral artery access, Outcome 4: Major complications (30‐day or in‐hospital)

Torsello 2003 reported two major complications in the cut‐down femoral artery access group and four major complications in the percutaneous group. In one case in the percutaneous group, the PS device was unsuccessful in closing the arterial entry site, and conversion to an open groin incision was necessary. In a further two cases, surgical repair of the access artery was necessary due to bleeding in one, and arterial thrombosis in the other. Two participants (one in each group) developed postoperative femoral artery occlusion requiring surgical thrombectomy. One participant in the cut‐down femoral artery access group required surgical repair of the access artery due to arterial thrombosis.

Nelson 2014 reported "unsuccessful treatment" due to procedural technical failure, a major adverse event or a vascular complication in 17 participants in the endovascular group and 11 participants in the cut‐down femoral artery access group. The number of major complications exceeded the number of participants with major complications, indicating some participants had more than one major complication.

In the cut‐down femoral artery access group, 18 major complications occurred (11 participants): neurological complications (three), renal failure (one), respiratory complications (one), secondary procedures (two), femoral neuropathy (one), haemorrhage (four), thrombosis (three), vascular injury (two), and endovascular device failure (one). The one endovascular device failure required switching to a lower‐profile Endologix system for EVAR and external iliac artery stenting. In the percutaneous group, 33 major complications occurred (17 participants): cardiac death (one), cardiac morbidity (two), renal failure (three), respiratory complications (three), secondary procedure (one), haemorrhage (six), thrombosis (four), vascular injury (four) and procedural technical failures (nine). There were nine procedural technical failures: ipsilateral preclose failures requiring surgical cut‐down/return to the operating theatre (eight) and thrombectomy/iliac stenting for thrombosis (one).

Vierhout 2019 reported 11 major complications in the percutaneous access group and five major complications in the open access group. In the percutaneous access group: conversions to open access because of percutaneous closure device failure (six), haematomas (two), pseudoaneurysms (two) and ARVI (one). In the open access group, there were SSIs (two), neuropathy (one) and haematomas (two).

Medium‐ to long‐term (six and 12 months) complications and mortality

One study reported complications and mortality at six months (Nelson 2014). There was no clear difference between groups (RR 0.82, 95% CI 0.25 to 2.65; P = 0.74; 135 participants; moderate‐certainty evidence; Analysis 1.5) (note the study authors did not report the P value; calculated by review authors). Between 31 and 210 days, seven participants in the percutaneous group and four in the cut‐down femoral artery access group had major adverse events or vascular complications. The number of complications exceeded the number of participants experiencing complications, indicating some participants had more than one complication. In the percutaneous group, there were nine major adverse events or vascular complications (seven participants): renal failure (one), ankle‐brachial index decrease (not requiring intervention) (three), cancer death (one), critical renal artery stenosis requiring stenting (one), and endoleak requiring secondary procedure (three). In the cut‐down femoral artery access group, there were four events in four participants: cancer death (one), endoleak requiring secondary procedure (two), and lymphocoele (one).

1.5. Analysis.

Comparison 1: Percutaneous versus cut‐down femoral artery access, Outcome 5: Medium‐ to long‐term complications (6 months)

In Nelson 2014, loss to follow‐up at six months was 10%. Loss to follow‐up was higher in the cut‐down femoral artery access group (7/50, 14%) than in the totally percutaneous group (8/100, 8%; note only 100 participants as one participant died in hospital and was, therefore, not included in the six‐month follow‐up data). Loss to follow‐up was for the following reasons in the totally percutaneous group: participant withdrawn (six) and participant missed follow‐up visit (two). In the cut‐down femoral artery access group loss to follow‐up was due to: participant withdrawn (three), participant missed follow‐up visit (two) and participant refused follow‐up (two). The same proportion of participants in each group withdrew (6%). Twice as many participants were missed in the cut‐down femoral artery access group (2/50, 4%) compared with the totally percutaneous group (2/100, 2%) and a further two participants (2/50, 4%) of the cut‐down femoral artery access group refused six‐month follow‐up.

Torsello 2003 only included in‐hospital data and did not report any medium‐ to long‐term complications or mortality. Vierhout 2019 only reported postoperative complications (less than 30 days) and did not report any medium‐ to long‐term complications or mortality.

Bleeding complications and haematoma (30‐day or in‐hospital)

Bleeding complications

Two studies reported bleeding complications (Nelson 2014; Torsello 2003). There was no clear difference in bleeding complications between the percutaneous and cut‐down femoral artery access groups (RR 1.02, 95% CI 0.29 to 3.64; P = 0.97; 181 participants; moderate‐certainty evidence; Analysis 1.6). There was no heterogeneity between the studies (I² = 0%).

1.6. Analysis.

Comparison 1: Percutaneous versus cut‐down femoral artery access, Outcome 6: Bleeding complications (30‐day or in‐hospital)

Torsello 2003 reported one bleeding‐related complication requiring surgical repair of the access artery in the percutaneous group. There was no difference between groups regarding haemoglobin loss (percutaneous group: 2 (SD 0.7) g/100 mL; cut‐down group: 2.2 (SD 0.9) g/100 mL) and haematocrit loss (percutaneous group: 6% (SD 2.5%); cut‐down group: 6.5% (SD 2.1%)). There were no blood transfusions in either group and they did not provide data on haematoma.

Nelson 2014 reported no haematomas in either group. There were six cases of bleeding in five participants in the percutaneous group, and four cases of bleeding in three participants in the cut‐down femoral artery access group. There were no clear differences in estimated blood loss between groups (cut‐down femoral artery access group: 280 (SD 290) mL; percutaneous PG group: 213 (SD 205) mL; percutaneous PS group: 193 (SD 198) mL; percutaneous PG versus cut‐down femoral artery access P = 0.115; percutaneous PS versus cut‐down femoral artery access P = 0.083). There were no clear differences in blood transfusion requirements between groups (cut‐down femoral artery access group: 14; percutaneous PG group: 4; percutaneous PS group: 16; percutaneous PG versus cut‐down femoral artery access P = 0.187; percutaneous PS versus cut‐down femoral artery access P = 1.000).

Haematomas

Two studies reported haematomas (Nelson 2014; Vierhout 2019). There was no clear difference between the percutaneous and cut‐down femoral artery access groups (RR 0.88, 95% CI 0.13 to 6.05; P = 0.89; 288 participants; Analysis 1.7). Note that although Analysis 1.7 considered two studies, Nelson 2014 reported no haematomas and, therefore, did not contribute to the effect estimate in this analysis.

1.7. Analysis.

Comparison 1: Percutaneous versus cut‐down femoral artery access, Outcome 7: Haematoma (30‐day or in‐hospital)

Operating time

Three studies reported operating time (Nelson 2014; Torsello 2003; Vierhout 2019). There was a difference in operating time between the percutaneous and cut‐down femoral artery access groups favouring percutaneous access (MD −21.13 minutes, 95% CI −41.74 to −0.53; P = 0.04; 318 participants; low‐certainty evidence; Analysis 1.8). There was substantial heterogeneity between the studies (I² = 73%). To facilitate meta‐analysis, we combined the results of the two different percutaneous groups (PS, PG) from Nelson 2014 (100.9 (SD 40.5) minutes) using the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022).

1.8. Analysis.

Comparison 1: Percutaneous versus cut‐down femoral artery access, Outcome 8: Operating time (minutes)

Torsello 2003 reported a difference in operating time (percutaneous group: 86.7 (SD 27) minutes; cut‐down femoral artery access group: 107.8 (SD 38.5) minutes; P < 0.05). Nelson 2014 reported a clear difference in operating time (percutaneous PG group: 107 (SD 45) minutes; cut‐down femoral artery access group: 141 (SD 73) minutes; P = 0.006; percutaneous PS group: 95 (SD 35) minutes; cut‐down femoral artery access group: 141 (SD 73) minutes; P < 0.01). Vierhout 2019 reported no clear difference in the duration of surgery (percutaneous group: 118 (SD 33) minutes; open surgical access: 125 (SD 32) minutes; P value not reported by study authors, calculated by review authors as P = 0.21).

Duration of intensive treatment unit stay

One study reported duration of ITU stay (Nelson 2014). There was no difference in length of ITU stay between groups (cut‐down femoral artery access group: 35 (SD 38) hours; percutaneous PG group: 26 (SD 9) hours; percutaneous PS group: 31 (SD 15) hours; percutaneous PG versus cut‐down femoral artery access P = 0.269; percutaneous PS versus cut‐down femoral artery access P = 0.614). Combination of the results of the two different percutaneous groups (PS, PG) gave a mean of (28.5 (SD 12.59) hours) using the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022). There was no clear difference in the duration of ITU stay between the percutaneous and cut‐down femoral artery access groups (MD −6.50 hours, 95% CI −17.32 to 4.32; P = 0.24; 151 participants; Analysis 1.9).

1.9. Analysis.

Comparison 1: Percutaneous versus cut‐down femoral artery access, Outcome 9: Duration of intensive treatment unit stay (hours)

Duration of hospital stay

One study reported duration of hospital stay (Nelson 2014). There was no difference in time to hospital discharge between groups (cut‐down femoral artery access group: 1.8 (SD 2.4) days; percutaneous PG group: 1.3 (SD 0.7) days; percutaneous PS group: 1.4 (SD 0.9) days; percutaneous PG versus cut‐down femoral artery access P = 0.135; percutaneous PS versus cut‐down femoral artery access P = 0.213). Combination of the results of the two different percutaneous groups (PS, PG) gave a mean of 1.4 (SD 0.8) days using the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022). There was no clear difference in the duration of hospital stay between the percutaneous and cut‐down femoral artery access groups (MD −0.4 days, 95% CI −1.08 to 0.28; P = 0.25; 151 participants; Analysis 1.10).

1.10. Analysis.

Comparison 1: Percutaneous versus cut‐down femoral artery access, Outcome 10: Duration of hospital stay (days)

Discussion

Summary of main results

This review compared the clinical outcomes of percutaneous access with cut‐down femoral artery access in elective bifurcated abdominal EVAR. With regard to the primary outcomes of mortality and failure of aneurysm exclusion, at least equivalence, but not necessarily superiority, over cut‐down EVAR had to be demonstrated to allow a recommendation, as percutaneous interventions may still be preferable if there are fewer wound problems and reduced length of hospital stay.

The three included studies recorded only one operative mortality and one endoleak requiring repair in the 30‐day follow‐up (Nelson 2014; Torsello 2003; Vierhout 2019). The results of the meta‐analyses indicated no clear difference in wound infections, 30‐day major complications, bleeding or haematomas across the studies. Six‐month data from one study, continued to show no clear difference in complications and mortality between percutaneous and cut‐down femoral artery access groups (Nelson 2014).

We detected a difference in operating time, with percutaneous access taking less time than cut‐down femoral artery access, with a shorter operating time representing a potential benefit.

Nelson 2014 reported no clear difference in ITU stay between groups. They found no difference between percutaneous and cut‐down femoral artery access groups in time from procedure to hospital discharge (percutaneous PG versus cut‐down femoral artery access P = 0.135; percutaneous PS versus cut‐down femoral artery access P = 0.213). This was not one of the outcomes selected by this Cochrane Review and we could not confirm its validity given the age of the study. However, this is in contrast to previous hypotheses that percutaneous access may reduce the length of hospital stay by proxy of fewer wound complications.

Overall completeness and applicability of evidence

The biggest limitation of this review is that it included only three studies, one with small sample size. This gave a pool of 318 participants to examine, with 151, 137 and 30 participants. Indeed, due to the small sample size (30 participants), Torsello 2003 might not be sufficiently powered to assess certain outcomes. Clearly, for a systematic review, this is a low number of studies and a limited pool of participants, which limits the widespread applicability of the results. There was a clear difference in the participant inclusion criteria between the studies. Torsello 2003 included participants with calcification of the femoral artery, scars in the groin or obesity, which are considered risk factors for wound complications and may make percutaneous access more challenging. By contrast, Vierhout 2019 and Nelson 2014 were more selective in their participants, specifically excluding morbidly obese people and those with significant scarring at the arterial access site or previous CFA surgery. Torsello 2003 reported insufficient data regarding the comorbidities (e.g. obesity) of the participants to examine this further. This may have influenced the outcomes included; however, overall this was not deemed significant as all other aspects of the trial conduction were comparable. Also of note were the differing sheath diameters used in the studies. Torsello 2003 and Vierhout 2019 reported using multiple sheath sizes, while Nelson 2014 reported using only 21F sheaths. This again may influence outcomes and may not be reflective of modern practice, where lower profile sheaths can be found.

None of the studies measured medium‐ to long‐term complications and mortality at 12 months. Only one study reported six‐month data for medium‐ to long‐term complications and mortality (Nelson 2014). For other outcomes (e.g. short‐term mortality, failure of aneurysm exclusion, bleeding complications, haematoma), only two studies contributed data. For the outcomes of duration of ITU stay and hospital stay, only one study reported data (Nelson 2014).

Torsello 2003 reported on the costs of the percutaneous and cut‐down femoral artery access interventions. The study investigators identified a higher cost of percutaneous interventions (EUR 474.7 (SD 109.7)) versus cut‐down femoral artery access (EUR 375.5 (SD 153.3); P < 0.01), which was, in their view, related to the cost of closure devices. The other two studies did not report costs (Nelson 2014; Vierhout 2019).

It was not within the scope of this update to investigate early mobilisation/return to normal activities or pain/analgesic use. However, as these are important factors, they should be thoroughly considered in future updates. The included studies reported limited information on these outcomes and we have reported this in Table 2. We did not undertake any quality assessment or meta‐analysis on these data.

1. Additional outcomes reported by trials.

Study ID	Early mobilisation/return to normal activities	Pain score/analgesic requirements
Nelson 2014	Time to ambulation (hours) No difference in time to ambulation (hours, PE 19 (SD 16), PEVAR/PG 17 (SD 7.2), PEVAR/PS 16 (SD 9.1); PEVAR/PG vs FE P = 0.388, PEVAR/PS vs FE P = 0.256)	Medication for groin pain (procedural/in‐hospital outcome) PE 30/50, PEVAR/PG 18/50, PEVAR/PS 12/51; PEVAR/PG vs FE P = 0.241, PEVAR/PS vs FE P = 0.029; fewer PEVAR participants prescribed analgesics for groin pain vs FE (15% vs 30%)
	Time to normal diet (hours) No difference in time to normal diet (hours, PE 15 (SD 22), PEVAR/PG 14 (SD 9.4), PEVAR/PS 10 (SD 8.4); PEVAR/PG vs FE P = 0.728, PEVAR/PS vs FE P = 0.135)	Ipsilateral groin pain (10‐point VAS was measured at predischarge, 1 month and 6 months) Only P values between groups were provided: Predischarge: PEVAR/PG vs FE P = 0.266, PEVAR/PS vs FE P = 0.009; 1 month: PEVAR/PG vs FE P = 0.423, PEVAR/PS vs FE P = 0.027; 6 months: PEVAR/PG vs FE P = 0.280, PEVAR/PS vs FE P = 0.656.
Torsello 2003	Time to ambulation (hours) A difference in time to ambulation (percutaneous 20.1 (SD 4.3) vs cut‐down 33.1 (SD 19.4); P < 0.001)	Not reported
Vierhout 2019	Not reported	VAS scores (measured at day 1 and 2 weeks postoperatively) The first day after intervention, a difference in VAS scores of 0.77 (95% CI 0.32 to 1.22; P = 0.001) was reported, in favour of the percutaneous access technique. Adjusted analysis showed a similar result of 0.69‐point difference in VAS (95% CI 0.27 to 1.12; P = 0.001).

CI: confidence interval; FE: femoral exposure; PEVAR: percutaneous endovascular aortic aneurysm repair; PG: 8F Perclose ProGlide device; PS: 10F Prostar XL device; SD: standard deviation; VAS: visual analogue scale.

This review identified no studies meeting inclusion criteria that considered the fascia suture technique, which has been reported as a potential alternative to closure devices for percutaneous access (Larzon 2006; Larzon 2015). We assessed two studies identified as ongoing in the second version of this review (Gimzewska 2017); one was included in Vierhout 2019 and the other was excluded due to including the wrong population (Uhlmann 2018). We identified no new ongoing studies.

Quality of the evidence

Of the three studies, Vierhout 2019 had a relatively large sample with few sources of bias except for a high risk of selective reporting (i.e. it did not report the primary outcome at one year as defined in their protocol). Nelson 2014 was methodologically stronger, had a relatively large sample size, good follow‐up and reported predefined outcome data with minimal risk of bias. Torsello 2003 had more limitations, namely a failure to adequately report the method of randomisation, allocation concealment and the preselected outcomes. Although in general it was a study of moderate certainty with complete follow‐up of all included participants, it remained hampered by its small sample size (30 participants).

We used the GRADEpro GDT software to assess the overall certainty of evidence in relation to each review outcome (GRADEpro GDT). We graded short‐term mortality as low‐certainty evidence by downgrading two evidence levels due to a low number of participants and events, and extremely wide CIs that included both harm and benefit (95% CI 0.06 to 36.18). We graded failure of aneurysm exclusion, wound infection, major complications, medium‐ to long‐term complications and bleeding complications as moderate‐certainty evidence, downgrading one evidence level due to a low number of participants and events, and a wide CIs that included both harm and benefit. Notably, the data for the failure of aneurysm exclusion and medium‐ to long‐term complications was available from only one study (Nelson 2014). For medium‐ to long‐term complications, although there were losses to follow‐up at six months, these were well described and there was no clear difference in the rate of loss between the percutaneous and cut‐down access groups. We downgraded operating time to low certainty, as the included studies were inadequately powered for this outcome and showed substantial heterogeneity.

Potential biases in the review process

The bias within the review process was minimised as far as possible. The most important aspect of this was a comprehensive literature search, encompassing grey literature as well as published sources. Since we considered all device types for totally percutaneous endovascular repairs and compared this against cut‐down femoral artery access for endovascular repair, we combined the data of the two different percutaneous groups (PS, PG) from Nelson 2014. For the continuous outcomes (e.g. operating time, duration of ITU stay and hospital stay), we used the methods described in the Cochrane Handbook for Systematic Reviews of Interventions to combine the data (Deeks 2022). We are aware that outcomes considered clinically relevant today may not have been identified in the original protocol (Jackson 2012), such as participant‐reported outcomes (e.g. health‐related quality of life, pain scores). We suggest that these are considered should new studies be identified in future updates.

Agreements and disagreements with other studies or reviews

One recent review examined four randomised controlled trials and found no difference in perioperative mortality, access site complications or infection, bleeding complications and haematoma, access‐related arterial injury, femoral artery occlusion, pseudo‐aneurysm and duration of hospital stay between the percutaneous and cut‐down EVAR (Antoniou 2021). In addition, Antoniou 2021 reported a shorter procedure time in the percutaneous EVAR group (MD –11.53 minutes, 95% CI –15.71 to –7.34). They graded the level of evidence as low or very low due to high risks of bias, few events and wide CIs. This broadly agrees with the findings of our review in terms of safety and efficacy but we downgraded the evidence by only one level because there were no concerns about risk of bias for most outcomes. Another difference was that Antoniou 2021 used the revised Cochrane RoB 2 tool for randomised trials to assess included studies for risk of bias (Sterne 2019). This will be considered for the next update. In addition, three RCTs overlapped in both reviews while Antoniou 2021 included one study that we excluded from our Cochrane Review as it did not meet our inclusion criteria (wrong population; Uhlmann 2018).

Another recent review included 13 studies (four randomised and nine retrospective studies) and found percutaneous access had a lower risk of overall complications, a shorter operation time and a higher technical success rate compared with the cut‐down access group (Bi 2022). In addition, Bi 2022 found no clear differences in other specific complications, such as wound infection, femoral artery thrombosis, bleeding complications, haematoma, nerve injury and dissection. For most outcomes, such as operation time and some specific complications, Bi 2022 and our Cochrane Review were consistent. However, integrating randomised controlled trials and retrospective studies in meta‐analyses is discouraged due to the methodological heterogeneity in the study design.

Another previous review examined 22 studies (one randomised, 10 non‐randomised and 11 retrospective studies) (Malkawi 2010). The review concluded that the percutaneous approach appeared safe and effective with low access‐related complications, but that further research was required to identify suitable candidates for the percutaneous approach. This broadly matches with the findings of our Cochrane Review with regard to safety and efficacy.

Due to the nature of our review (only including randomised controlled trials), the pool of data for this review was substantially smaller than the other reviews mentioned here and, as such, we reached no firm conclusions. We found no changes in the conclusion since the previous version of this review, which was published in 2017 (Gimzewska 2017).

Authors' conclusions

Implications for practice.

Skin puncture may make little to no difference to short‐term mortality. There is probably little or no difference in failure of aneurysm exclusion (failure to seal the aneurysms), wound infection, major complications within 30 days or while in hospital, medium‐ to long‐term (six months) complications and bleeding complications between the two groups. Compared with exposing the femoral artery, skin puncture may reduce the operating time slightly.

Implications for research.

This review has identified a need for further research into a potentially beneficial technique. Further independent and sufficiently powered trials, such as a robust and sufficiently powered multicentre randomised trial, would also be helpful. The authors of future research and reviews may wish to examine a number of the areas identified within this review, namely differences in cost, operating time, duration of hospital and intensive treatment unit stay, wound infections, potential bleeding and haematoma complications, and operative success. It is also important that future trials investigate early mobilisation/return to normal activities and pain/analgesic requirements and explore the impact of basic patient characteristics (e.g. obesity). Measuring these, along with core issues such as mortality, could give a more complete picture of the possible benefits and drawbacks of the differing approaches to endovascular aneurysm repair and more fully inform future surgical decisions. In addition, if the technique is found to be effective in elective aneurysm repairs it may be considered in the future for the treatment of ruptured abdominal aortic aneurysms.

What's new

Date	Event	Description
8 September 2022	New search has been performed	New search run. One new included study and six new excluded studies identified.
8 September 2022	New citation required but conclusions have not changed	New search run. One new included study and six new excluded studies identified. New review team has updated the review. Text amended to reflect Cochrane standards. 'Summary of findings' table updated. Conclusions not changed.

History

Protocol first published: Issue 11, 2012
Review first published: Issue 2, 2014

Date	Event	Description
26 October 2016	New search has been performed	New search run. One new included study, one new excluded study and two new ongoing studies were identified.
26 October 2016	New citation required but conclusions have not changed	New search run. One new included study, one new excluded study and two new ongoing studies were identified. New author joined team. Text amended to reflect Cochrane standards. 'Summary of findings' table added. Conclusions not changed.

Appendices

Appendix 1. Search strategies

Source	Search strategy	Hits retrieved
1. VASCULAR REGISTER IN CRSW (Date of most recent search: 8 April 2022)	#1 percutaneous* AND INREGISTER #2 PEVAR AND INREGISTER #3 preclose AND INREGISTER #4 ProGlide AND INREGISTER #5 Prostar XL AND INREGISTER #6 Perclose AND INREGISTER #7 #1 OR #2 OR #3 OR #4 OR #5 OR #6 #8 aneurysm* AND INREGISTER #9 aneurIsm* AND INREGISTER #10 AAA AND INREGISTER #11 #8 OR #9 OR #10 #12 #11 AND #7	April 2022: 49
2. CENTRAL via CRSO (Date of most recent search: 8 April 2022)	#1 MESH DESCRIPTOR Aortic Aneurysm EXPLODE ALL TREES 801 #2 MESH DESCRIPTOR Aneurysm, Ruptured EXPLODE ALL TREES 197 #3 MESH DESCRIPTOR Aorta, Abdominal EXPLODE ALL TREES 338 #4 AAA*:TI,AB,KY 1195 #5 (aneurysm* adj4 (abdom* or thoracoabdom* or thoraco‐abdom* or aort*)):TI,AB,KY 994 #6 (aort* adj3 (balloon* or dilat* or bulg*)):TI,AB,KY 720 #7 (abdom* adj3 (balloon* or dilat* or bulg*)):TI,AB,KY 54 #8 (aneurism* adj4 (abdom* or thoracoabdom* or thoraco‐abdom* or aort*)):TI,AB,KY 4 #9 #1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 3002 #10 percutaneous*:TI,AB,KY 21244 #11 PEVAR:TI,AB,KY 14 #12 preclose:TI,AB,KY 5 #13 ProGlide:TI,AB,KY 40 #14 (Prostar XL):TI,AB,KY 9 #15 Perclose:TI,AB,KY 33 #16 #10 OR #11 OR #12 OR #13 OR #14 OR #15 21268 #17 #9 AND #16 216	April 2022: 216
3. ClinicalTrials.gov (Date of most recent search: 8 April 2022)	aneurysm* OR aneurIsm* OR AAA | percutaneous OR PEVAR OR preclose OR ProGlide OR Prostar OR Perclose	April 2022: 14
4. ICTRP Search Portal (Database not available 8 April 2022)	aneurysm* OR aneurIsm* OR AAA | percutaneous OR PEVAR OR preclose OR ProGlide OR Prostar OR Perclose	April 2022: 0
5. MEDLINE (Ovid MEDLINE Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE Daily and Ovid MEDLINE) 1946 to present (Date of most recent search: 8 April 2022)	1 exp Aortic Aneurysm/ 2 exp Aneurysm, Ruptured/ 3 exp Aorta, Abdominal/ 4 AAA*.ti,ab. 5 (aneurysm* adj4 (abdom* or thoracoabdom* or thoraco‐abdom* or aort*)).ti,ab. 6 (aort* adj3 (balloon* or dilat* or bulg*)).ti,ab. 7 (abdom* adj3 (balloon* or dilat* or bulg*)).ti,ab. 8 (aneurism* adj4 (abdom* or thoracoabdom* or thoraco‐abdom* or aort*)).ti,ab. 9 or/1‐8 10 percutaneous*.ti,ab. 11 PEVAR.ti,ab. 12 preclose.ti,ab. 13 ProGlide.ti,ab. 14 Prostar XL.ti,ab. 15 Perclose.ti,ab. 16 or/10‐15 17 9 and 16 18 randomized controlled trial.pt. 19 controlled clinical trial.pt. 20 randomized.ab. 21 placebo.ab. 22 drug therapy.fs. 23 randomly.ab. 24 trial.ab. 25 groups.ab. 26 or/18‐25 27 exp animals/ not humans.sh. 28 26 not 27 29 17 and 28	April 2022: 552
6. Embase via Ovid (Date of most recent search: 8 April 2022)	1 exp aortic aneurysm/ 2 exp aneurysm rupture/ 3 exp abdominal aorta/ 4 AAA*.ti,ab. 5 (aneurysm* adj4 (abdom* or thoracoabdom* or thoraco‐abdom* or aort*)).ti,ab. 6 (aort* adj3 (balloon* or dilat* or bulg*)).ti,ab. 7 (abdom* adj3 (balloon* or dilat* or bulg*)).ti,ab. 8 (aneurism* adj4 (abdom* or thoracoabdom* or thoraco‐abdom* or aort*)).ti,ab. 9 or/1‐8 10 percutaneous*.ti,ab. 11 PEVAR.ti,ab. 12 preclose.ti,ab. 13 ProGlide.ti,ab. 14 Prostar XL.ti,ab. 15 Perclose.ti,ab. 16 or/10‐15 17 9 and 16 18 randomized controlled trial/ 19 controlled clinical trial/ 20 random$.ti,ab. 21 randomization/ 22 intermethod comparison/ 23 placebo.ti,ab. 24 (compare or compared or comparison).ti. 25 ((evaluated or evaluate or evaluating or assessed or assess) and (compare or compared or comparing or comparison)).ab. 26 (open adj label).ti,ab. 27 ((double or single or doubly or singly) adj (blind or blinded or blindly)).ti,ab. 28 double blind procedure/ 29 parallel group$1.ti,ab. 30 (crossover or cross over).ti,ab. 31 ((assign$ or match or matched or allocation) adj5 (alternate or group$1 or intervention$1 or patient$1 or subject$1 or participant$1)).ti,ab. 32 (assigned or allocated).ti,ab. 33 (controlled adj7 (study or design or trial)).ti,ab. 34 (volunteer or volunteers).ti,ab. 35 trial.ti. 36 or/18‐35 37 17 and 36	April 2022: 1025
7. CINAHL via EBSCO (Date of most recent search: 8 April 2022)	S30 S16 AND S29 S29 S17 OR S18 OR S19 OR S20 OR S21 OR S22 OR S23 OR S24 OR S25 OR S26 OR S27 OR S28 S28 MH "Random Assignment" S27 MH "Triple‐Blind Studies" S26 MH "Double‐Blind Studies" S25 MH "Single‐Blind Studies" S24 MH "Crossover Design" S23 MH "Factorial Design" S22 MH "Placebos" S21 MH "Clinical Trials" S20 AB placebo* S19 TX random* S18 TX trial* S17 TX "latin square" S16 S8 AND S15 S15 S9 OR S10 OR S11 OR S12 OR S13 OR S14 S14 TX Perclose S13 TX Prostar XL S12 TX ProGlide S11 TX preclose S10 TX PEVAR S9 TX percutaneous* S8 S1 OR S2 OR S3 OR S4 OR S5 OR S6 OR S7 S7 TX (aneurism* n4 (abdom* or thoracoabdom* or thoraco‐abdom* or aort*)) S6 TX (abdom* n3 (balloon* or dilat* or bulg*)) S5 TX (aort* n3 (balloon* or dilat* or bulg*)) S4 TX (aneurysm* n4 (abdom* or thoracoabdom* or thoraco‐abdom* or aort*)) S3 TX AAA* S2 (MH "Aorta, Abdominal") S1 (MH "Aortic Aneurysm+")	April 2022: 147
TOTAL before deduplication		2003
TOTAL after deduplication		1462

Data and analyses

Comparison 1. Percutaneous versus cut‐down femoral artery access.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Short‐term mortality (30‐day or in‐hospital)	2	181	Risk Ratio (M‐H, Random, 95% CI)	1.50 [0.06, 36.18]
1.2 Failure of aneurysm exclusion	1	151	Risk Ratio (M‐H, Random, 95% CI)	0.17 [0.01, 4.02]
1.3 Wound infection (30‐day or in‐hospital)	3	318	Risk Ratio (M‐H, Random, 95% CI)	0.18 [0.01, 3.59]
1.4 Major complications (30‐day or in‐hospital)	3	318	Risk Ratio (M‐H, Random, 95% CI)	1.21 [0.61, 2.41]
1.5 Medium‐ to long‐term complications (6 months)	1	135	Risk Ratio (M‐H, Random, 95% CI)	0.82 [0.25, 2.65]
1.6 Bleeding complications (30‐day or in‐hospital)	2	181	Risk Ratio (M‐H, Random, 95% CI)	1.02 [0.29, 3.64]
1.7 Haematoma (30‐day or in‐hospital)	2	288	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.13, 6.05]
1.8 Operating time (minutes)	3	318	Mean Difference (IV, Random, 95% CI)	‐21.13 [‐41.74, ‐0.53]
1.9 Duration of intensive treatment unit stay (hours)	1	151	Mean Difference (IV, Random, 95% CI)	‐6.50 [‐17.32, 4.32]
1.10 Duration of hospital stay (days)	1	151	Mean Difference (IV, Random, 95% CI)	‐0.40 [‐1.08, 0.28]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Nelson 2014.

Study characteristics
Methods	Study design: RCT Method of randomisation: block
Participants	Country: USA (18 centres) Number: 151 Age (years): PG group: 70 (SD 6.6); PS group: 74 (SD 11); femoral cut‐down (cut‐down femoral artery access) group: 73 (SD 8.8) Sex: 136 men, 15 women BMI (kg/m2): PG group: 29 (SD 3.9); PS: 28 (SD 4.7); femoral cut‐down group: 28 (SD 4.7) Inclusion criteria: presenting with AAA with maximal diameter ≥ 5 cm or rapidly expanding aged > 18 years informed consent form understood and signed, and agreement to all follow‐up visits suitable ipsilateral CFA for percutaneous access with preclose technique – with criteria for this specified anatomical eligibility for Endologix System (EVAR stent) life expectancy > 1 year (judged by the investigator) serum creatinine level ≤ 1.7 mg/dL no planned major intervention or surgery within 30 days after study procedure not morbidly obese (BMI < 40 kg/m2) Exclusion criteria: prior groin incision, haematoma or significant scarring at the ipsilateral arterial access site, clip‐based vascular closure device placement ever, collagen‐based vascular closure device placement in either arterial access site within prior 90 days, femoral artery needle puncture in either arterial access site within prior 30 days localised groin infection, traumatic vascular injury, femoral artery aneurysm, arteriovenous fistula, pseudoaneurysm allergy to any device component, coagulopathy or uncontrolled bleeding disorder, active systemic infection, connective tissue disease, prior renal transplant cerebrovascular event or myocardial infarction within 3 months of enrolment
Interventions	All aneurysms were repaired using the Endologix 21F profile‐based sheath system (Endologix, Inc, Irvine, California, US). Open FE, 50 participants. Small oblique incision made, with direct exposure and control of the femoral artery. Access site closure achieved with direct artery repair and layered wound closure. PG device, 50 participants. PG device (Abbott Vascular, Inc, Redwood City, California, US) used in the preclose technique for closure of the access site. PS, 51 participants. PS device (Abbott Vascular) used in the preclose technique for closure of the access site. Postdeployment graft position/aneurysm exclusion confirmed angiographically in all cases.
Outcomes	Primary endpoint: 'treatment success'; a composite endpoint comprising procedural technical success and absence of major adverse events or vascular complications at 30 days. Secondary endpoints: activated clotting time, contrast volume, fluoroscopy time, estimated blood loss, blood transfusion, procedure time, ipsilateral time to haemostasis, time to ambulation, time to normal diet, ICU length of stay, medication for groin pain and time to hospital discharge.
Funding	Funded by Endologix, Inc, with financial contribution by Abbott Vascular, Inc, and Abbott Vascular, Inc provided ipsilateral closure devices free of charge.
Declarations of interest	1 study author served on the Scientific Advisory Board for Endologix, Inc, and 3 authors are on the Endologix, Inc speaker's bureau.
Notes	2 further manuscripts presented the 30‐day results of the same trial (Krajcer 2013; Nelson 2012). The data extracted from Krajcer 2013 and Nelson 2012 were the same as Nelson 2014 and were not considered as a separate data set, which was cross‐checked to ensure reporting consistency. Another report included as a report for checking its registration information (NCT01070069).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation by study site, using 2 block sizes (3 or 6) with random choice of block size order. Sealed randomisation envelopes were used. 2:1 randomisation (PEVAR:FE) was used, with equal allocation to 2 PEVAR groups (PG, PS).
Allocation concealment (selection bias)	Low risk	Sealed randomisation envelopes were used: on screening eligibility confirmation the next sequential randomisation envelope was opened and assignment was immediately allocated.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Blinding was not possible due to nature of surgical interventions.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Unlikely relating to the primary endpoint as they were objective in nature.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants reported no loss to follow‐up at 30 days. There was a 10% loss to follow‐up at 6 months. This only affected 1 secondary outcome in the review (medium‐ to long‐term complications). Reasons for loss to follow‐up were given for each group.
Selective reporting (reporting bias)	Low risk	The primary endpoint, secondary outcomes and subgroup analysis were clearly defined and reported.
Other bias	Low risk	None identified.

Torsello 2003.

Study characteristics
Methods	Study design: RCT Method of randomisation: not reported
Participants	Country: Germany Number: 30 Age (years): 51–90 (mean 72.9 (SD 9.9)) Sex: 29 men, 1 woman BMI (kg/m2): not reported (obesity: 6/15 participants for percutaneous group and 7/15 participants for cut‐down group) Inclusion criteria: presenting with an AAA, including those with calcification of femoral artery, scars in the groin or obesity Exclusion criteria: psychiatric conditions undergoing the implantation of an aorto‐uni‐iliac graft with an aneurysm of the femoral artery
Interventions	Aneurysms were repaired using either the Zenith graft (Cook, Bloomington, Indiana, US), 16 participants, or the Talent endovascular graft (Medtronic, Sunrise, Florida, US), 14 participants. Percutaneous access, 15 participants. A PS percutaneous vascular surgery device (Abbott Vascular) used in the preclose technique for closure of the access site. Surgical cut‐down (femoral artery access), 15 participants. A transverse groin incision made to expose the CFA for direct needle puncture. All participants received heparin 5000 IU after sheath insertion. Duplex ultrasound scanning performed before and after the procedure.
Outcomes	Outcomes: operative success (the successful closure of the femoral artery following insertion of graft); in‐hospital mortality and major complications; wound complications; re‐intervention rate; blood loss; operative time and time to ambulation; total operative cost.
Funding	Not reported.
Declarations of interest	None.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not reported.
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not reported.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Blinding was not possible due to nature of surgical interventions.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Unlikely given the outcome measures used were objective in nature.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants were fully reported with no loss to follow‐up due to small numbers and a short follow‐up timescale.
Selective reporting (reporting bias)	Unclear risk	Selected outcomes were not outlined, which made the study paper more prone to selective reporting; however, most major outcomes appeared to have been reported.
Other bias	Low risk	None identified.

Vierhout 2019.

Study characteristics
Methods	Study design: RCT Method of randomisation: using blinded envelopes
Participants	Country: the Netherlands Number: 137 Age (years): mean age 72 (range 56–93); percutaneous access group: 72.6 (SD 8.1); open surgical access group: 72.4 (SD 6.2) Sex: 123 men, 14 women BMI (kg/m2): percutaneous access group: 27.5 (SD 3.6); open surgical access group: 27.2 (SD 3.7) Inclusion criteria: suitable for an elective EVAR aged > 18 years physically and mentally capable of giving consent diagnosed with an AAA with diameter ≥ 5.5 cm, and growth of ≥ 5 mm in 6 months Exclusion criteria: CFA access sites with > 50% circumferential calcification (based on multidisciplinary discussion and a radiologist report) previous CFA surgery infection at time of operation quote "fewer or more than two access sites" (additional brachial or carotid) requiring fenestrated or branched EVAR
Interventions	Each participant with an AAA suitable for EVAR was operated via arterial access in both groins. The participants were randomised to either open or percutaneous access to the main device through the CFA. The access on the contralateral side was operated with the opposite technique. The type of endovascular stent graft was not reported. Percutaneous group (percutaneous access of main device), 73 participants. Access to main device obtained via an ultrasound‐guided puncture of the CFA, followed by the positioning of ≥ 1 PS or PG devices (Abbott Vascular) via an incision of approximately 1 cm. After completion of angiography and removal of sheaths, haemostasis was obtained by applying gentle manual pressure to the proximal CFA and advancing the preformed knot of the closure device with a knot pusher. Open group (open surgical access of main device), 64 participants. Access to main device obtained via a 3‐ to 4‐cm craniocaudal incision and puncture of the CFA under direct vision. After completion of angiography and removal of the sheaths, haemostasis was obtained using a running, interrupted or crossing polypropylene suture. All participants in both groups received prophylactic antibiotics 15 to 60 minutes before surgery and intravenous heparin 5000 units during surgery.
Outcomes	Primary endpoint: number of SSIs Secondary endpoints: wound complications, standardised wound assessment scores/Southampton Wound Assessment score, VAS score, access‐related vascular injuries, serious adverse events, duration of surgery and blood loss.
Funding	Abbott Vascular funded this research (reference number 311208). A database manager collected data from the participating centres. Members of the PiERO study group were not financially rewarded. Devices were delivered at usual costs.
Declarations of interest	None.
Notes	1 conference abstract presented 2‐week postoperative results, which were different from those in the final report (Vierhout 2016). Given the abstract was not peer‐reviewed and only presented the interim results, we did not consider it a separate data set. We considered the inconsistencies between reports in the assessment the risk of bias. Another manuscript was the protocol of this study and was used to check details of its design and methods and to assess the risk of bias (Vierhout 2015).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation performed using sealed envelopes prior to operation at each centre. Quote: "These envelopes were produced outside the centers and delivered in a 1:1 ratio and an amount of 20 envelopes per batch." Participants were randomised to the percutaneous or the open CFA access group in a 1:1 ratio.
Allocation concealment (selection bias)	Low risk	The operating team (operating physician or assistant) could not be blinded to the allocation and, therefore, drew a sealed envelope to allocate the participants to 1 of the 2 groups. The allocation was concealed to participants, follow‐up physicians, data collectors and data analysts.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The study was single‐blinded since the surgeon could not be blinded when deciding the access side for main device based on the participant's anatomy. The operations (the introduction side and the main device introduction technique) were concealed to participants.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The operations (the introduction side and the main device introduction technique) were concealed to follow‐up physicians, data collectors and data analysts. Also, the primary endpoint measurement was objective in nature and was assessed by an independent researcher who was blinded to the main device access site. Data were analysed anonymously.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All 137 assigned participants were fully reported with no loss to follow‐up at 6 weeks postoperatively.
Selective reporting (reporting bias)	High risk	In the protocol, it was stated that the primary endpoint (number of SSI) would be evaluated within 30 days postoperatively and 1 year after the operation. The data regarding the primary endpoint at 1 year were not reported in the full text. In addition, we found some inconsistencies in results between study reports.
Other bias	Low risk	None identified.

AAA: abdominal aortic aneurysm; BMI: body mass index; CFA: common femoral artery; EVAR: endovascular aneurysm repair; FE: femoral exposure; ICU: intensive care unit; PEVAR: percutaneous endovascular aortic aneurysm repair; PG: 8F Perclose ProGlide device; PS: 10F Prostar XL device; RCT: randomised controlled trial; SSI: surgical site infections; VAS: visual analogue scale.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Hattab 2012	Did not include primary intervention. Study only examined percutaneous closure devices not aneurysm repairs. (Hattab 2012 was not identified by the search for this version of the review but was excluded in the previous version.)
Ichihashi 2016	Non‐randomised.
Iłżecki 2018	Non‐randomised.
Jean‐Baptiste 2008	Non‐randomised.
Krajcer 2010	Non‐randomised.
Krajcer 2011	Non‐randomised.
NTR4225	We contacted trial registries but we were unable to locate any further details.
Pitton 2003	Non‐randomised – determined using Google Translate/DeepL Translator and checked by a colleague who can read German (Lu Fan, PhD candidate, from Klinik für Hals‐, Nasen‐, Ohrenheilkunde, Medizinische Hochschule Hannover).
Roche‐Nagle 2018	Non‐randomised.
Uhlmann 2018	Wrong population: only 66% (infrarenal aortic aneurysms, 33 participants) of included EVAR indication participants had AAA and the results were not analysed based on the category of EVAR indication. Usually, 80% is the threshold for the percentage of the targeted population in the overall population.
Xiong 2012	Non‐randomised.

AAA: abdominal aortic aneurysm; EVAR: endovascular aneurysm repair.

Differences between protocol and review

2014 version

We added the outcome 'Bleeding complications and haematoma' to the original review.

2017 version

We added the outcome 'Duration of hospital stay' to the second version of the review.

2022 version

We used a random‐effects model instead of a fixed‐effect model because of potential heterogeneity (variation across studies) in the third version of the review.

We changed the outcome 'aneurysm exclusion' into 'failure of aneurysm exclusion' in the third version of the review to accurately reflect the data analysed.

We changed the outcome 'long‐term (12 months) complications and mortality' to 'medium‐ to long‐term (6 and 12 months) complications and mortality' in the third version of the review to include six‐month data.

We reported the outcome 'Bleeding complications and haematoma' and analysed them separately in the third version of the review.

Contributions of authors

QW: co‐ordinating the review within the team, screening search results, organising retrieval of papers, screening retrieved papers against inclusion criteria, abstracting data from papers, assessing the risk of bias, writing to registry platforms for additional information, data management for the review, entry of data, data analysis, GRADE assessment, data interpretation, writing the review.

JW: co‐ordinating the review with the editor, screening search results, screening retrieved papers against inclusion criteria, data analysis, data interpretation, writing the review.

YM: screening search results, screening retrieved papers against inclusion criteria, abstracting data from papers, assessing the risk of bias.

YZ: screening search results, screening retrieved papers against inclusion criteria, abstracting data from papers, assessing the risk of bias.

XS: screening search results, abstracting data from papers, assessing the risk of bias.

SX: assessing the risk of bias, data analysis, data interpretation.

FL: screening search results, data interpretation, GRADE assessment, providing clinical guidance on the whole process.

MG: critically reviewing and revising the updated manuscript.

ML: data interpretation, providing methodological guidance on the risk of bias and GRADE assessment.

LY: providing methodological guidance on the whole process.

All review authors critically reviewed the manuscript and approved the final version.

Sources of support

Internal sources

• No internal sources of support, Other

  No internal sources of support

External sources

• Chief Scientist Office, Scottish Government Health Directorates, The Scottish Government, UK

  The Cochrane Vascular editorial base is supported by the Chief Scientist Office.

Declarations of interest

QW: none.

JW: none

YM: none.

YZ: none.

XS: none.

SX: none.

FL: none.

MG: none.

ML: none.

LY: none.

New search for studies and content updated (no change to conclusions)