Systematic Review of Renal Outcomes and Procedural Efficacy of Carbon Dioxide Digital Subtraction Angiography in Endovascular Aortic Repair

Abstract

Objective

Carbon dioxide (CO₂) has been proposed as a contrast agent during endovascular aortic repair (EVAR) to minimise the risk of worsened post-operative renal function (PO-RFW). However, strong clinical evidence supporting its widespread use remains limited. This systematic review aimed to evaluate current literature on the use of CO₂ digital subtraction angiography (DSA) during EVAR, with particular attention to renal outcomes and procedural efficacy.

Data Sources

Systematic searches were conducted in PubMed, Scopus, and the Cochrane Library for studies published up to December 2024 reporting the use of CO₂-DSA during standard EVAR.

Review Methods

The review was registered prospectively in the International Prospective Register of Systematic Reviews (PROSPERO) database (CRD42024580889). All papers reporting the use of CO₂-DSA during EVAR were included. The primary outcome was the incidence of PO-RFW, and secondary outcomes included procedural success, arterial image visualisation quality, endoleak detection, and radiation exposure.

Results

Of the 108 articles initially identified, 16 studies published between 2007 and 2024 met the inclusion criteria, reporting data on 901 patients (mean age 76 years; 89.1% men) undergoing EVAR with CO₂-DSA. Among these, 622 (69.0% ) were treated exclusively with CO₂-DSA. All studies published since 2018 adopted an automatic injector. Seven studies comparedPO-RFW using different definitions; three of the included studies reported a statistically significantrole for CO₂-DSA in renal protection. Three studies reported a higher dose of radiation during CO₂-EVAR. Renal and hypogastric arteries were effectively visualised during repair in 53–100% and 94–100% of cases, respectively. Data on endoleak detection were heterogeneous, limiting definitive conclusions.

Conclusion

The use of CO₂ as contrast medium in EVAR appears to be safe and effective. Evidence suggests potential benefits for renal protection, although inconsistencies are reported regarding study design and outcome definitions. Further research is necessary to standardise CO₂ delivery, imaging assessment criteria, and renal outcome definitions. To standardise clinical practice, a defined protocol is needed to maximise its advantages in everyday practice.

Highlights

• •Systematic review on CO₂-DSA use during endovascular aortic repair (EVAR).
• •Sixteen studies (901 patients) included following PRISMA and PROSPERO standards.
• •CO₂-DSA used exclusively in 69% of EVAR cases with zero iodinated contrast.
• •Three studies demonstrated significant renal protection with CO₂-DSA.
• •Evidence supports safety; standardisation of CO₂ protocols remains essential.

INTRODUCTION

Endovascular aortic repair (EVAR) has revolutionised the management of infrarenal abdominal aortic aneurysms, offering a minimally invasive alternative to open surgical repair with lower peri-operative morbidity and mortality, particularly in high risk patients.¹

Despite these advantages, acute kidney injury (AKI) remains a frequent post-operative complication, with an incidence of up to 20% following EVAR.^2,3 The most common aetiology is contrast induced nephrotoxicity, particularly in patients with pre-existing chronic kidney disease (CKD).^4,5 This has been categorised as post-contrast AKI (PC-AKI), with a reported incidence ranging from 4% to 18%, and is often associated with the volume of iodinated contrast medium (ICM) administered during the procedure.6, 7, 8 To mitigate this risk, carbon dioxide (CO₂) has been proposed as an alternative contrast agent for digital subtraction angiography (DSA). In recent years, CO₂-DSA has been used increasingly in peripheral and aortic endovascular procedures, including both standard and complex EVAR.9, 10, 11 CO₂-DSA may be used either as the sole contrast agent or in combination with ICM, enabling high quality imaging while minimising renal exposure to iodine based agents.

Nevertheless, the clinical adoption of CO₂-DSA in EVAR remains limited. Evidence supporting its efficacy and safety is still emerging and is largely confined to single centre studies and small multicentre series, which nonetheless suggest a promising role for the technique.¹² Importantly, the most recent 2024 European Society for Vascular Surgery Clinical Practice Guidelines did not provide formal recommendations on the use of CO₂-DSA for the prevention of PC-AKI.¹³

In addition, potential technical limitations, including suboptimal visualisation of proximal and distal landing zones, and intra- and peri-procedural clinical complications remain areas of active investigation.^14,15

The aim of this systematic review was to critically evaluate the current literature on the use of CO₂-DSA during EVAR for infrarenal abdominal aortic aneurysms, with particular emphasis on procedural safety and technical considerations.

METHOD

Search strategy

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines¹⁶ and was registered a priori in the International Prospective Register of Systematic Reviews (PROSPERO) database (CRD42024580889). Two independent reviewers (P.S. and F.C.) systematically searched PubMed, Scopus, and the Cochrane Library for English language studies published up to December 2024 that reported the use of CO₂-DSA in standard EVAR. The search strategy was developed using the Population, Intervention, Comparison, and Outcome framework. The following search strategy was adopted: (((carbon dioxide) OR (carbon-dioxide)) OR (carbon-dioxide digital subtraction angiography)) AND (((endovascular aortic repair) OR (EVAR)) OR (endovascular abdominal aortic repair)). Rayyan software (Ryyan Systems Inc., Cambridge, MA, USA) was used to assist with title and abstract screening and full text review. Duplicate records were removed, and two reviewers independently screened titles, abstracts, and full texts according to pre-defined eligibility criteria. Disagreements were resolved by a third reviewer (E.G.), and the final inclusion set was confirmed by the senior author (M.G.).

Inclusion and exclusion criteria

Eligible studies included patients undergoing standard EVAR for aortic or iliac pathology with intra-operative CO₂-DSA. Studies using CO₂ as the sole contrast agent (so called ‘zero contrast’ procedure) or in combination with ICM were included. Both elective and emergency cases were considered. Comparative studies (CO₂ vs. ICM) and non-comparative series reporting outcomes following CO₂-DSA were included.

Studies were excluded if they involved open surgical repair, advanced endovascular techniques (e.g., fenestrated EVAR and branched EVAR), or series with less than five patients. Duplicate series by the same research group were excluded unless they involved distinct populations or methodologies.

Outcomes and definitions

The primary outcome was the rate of post-operative renal function worsening (PO-RFW). This included any author defined renal deterioration during the first 30 post-operative days, including changes in serum creatinine (sCr) or estimated glomerular filtration rate (eGFR). Pre-operative CKD was defined as an eGFR <60 mL/min/1.73 m².¹⁷

Secondary technical endpoints were intra-operative total radiation exposure, intra-operative fluoroscopy time, total procedure duration, accuracy of arterial visualisation to adequately define proximal and distal landing zones, and accuracy of intra-operative endoleak (EL) detection.

Data extraction and risk of bias assessment

Two reviewers (P.S. and F.C.) independently extracted study data from a standardised electronic database, collecting information on study design, setting, population, safety outcomes, and technical metrics. Risk of bias was assessed using the Risk Of Bias In Non-randomised Studies - of Interventions (ROBINS-I) for observational studies and RoB 2.0 for randomised trials. Risk of bias was categorised as low, moderate, or serious. Discrepancies were resolved by discussion with additional authors (E.G. and M.G.). RobVis tool (McGuinnes LA, Higgins JPT, Bristol, UK) was used for graphical display of bias assessments.

Synthesis of the results

A proportion meta-analysis was planned for the primary outcomes, where feasible. Otherwise, data were summarised descriptively and tabulated. Continuous variables were reported as mean ± standard deviation or median with interquartile range (quartiles 1–3), depending on data availability. Categorical variables were expressed as counts and percentages.

RESULTS

CO₂-DSA in EVAR

Of the 108 studies retrieved after duplicate removal and title and abstract screening, 28 were selected for full text review. Eventually, 16 studies^11,18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32 that reported on EVAR performed with CO₂-DSA were included (Fig. 1).

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart of the selection process in PubMed, Scopus, and the Cochrane Library for English language studies published up to December 2024 that reported the use of carbon dioxide digital subtraction angiography in standard endovascular aortic repair.

Characteristics of the included studies and the focus of each article are summarised in Table 1. Twelve retrospective, three prospective,^11,21,28 and one randomised controlled trial²³ were included. Two were multicentre cohort studies.^11,28 Overall, 901 patients were considered eligible (89.1% men; mean age 76 ± 3 years). They were all treated for abdominal aortic aneurysms with standard EVAR. Pre-operative CKD was present in 260 (28.8%) patients.

Table 1.

Studies included in the systematic review are listed in chronological order, with the number of carbon dioxide patients and the main details.

Author (year)	Journal	Study design	Centre(s)	Enrolment	CO2 patients – n∗	Zero contrast cases – n∗	Aim of the study
Chao et al.18 (2007)	JVS	R	Los Angeles, California, USA	2003–2005	16	3	Endoleak detection with CO2-DSA
Lee et al.32 (2010)	Vasc and Endovasc Surgery	R	Kingswood, Australia	2009	17	0	Evaluation of CO2-DSA in arterial visualisation and graft deployment
Huang et al.19 (2013)	Ann Vasc Surg	R	Los Angeles, California, USA	2006–2010	76	76	Endoleak detection with CO2-DSA
Criado et al.20 (2012)	JVS	R	Michigan, USA	2008–2011	114	72	Comparison of post-operative renal function outcomes between CO2 and CO2 + ICM angiographies
Sueyoshi et al.21 (2015)	JVS	P	Nagasaki, Japan	2011–2013	40	40	Endoleak detection with CO2-DSA
De Angelis et al.22 (2017)	Int J Cardiovas Imaging	R	Milano, Italy	2010–2015	13	12	Evaluation of CO2-DSA in arterial visualisation and graft deployment
De Almeida Mendes et al.23 (2017)	Ann Vasc Surg	RCT	São Paulo, Brazil	2012–2014	16	6	Evaluation of post-operative renal function outcomes with CO2-DSA
Takeuchi et al.24 (2018)	Ann Vasc Surg	R	Yamaguchi, Japan	2012–2016	30	0	Evaluation of post-operative renal function outcomes with CO2-DSA
Mascoli et al.25 (2018)	Ann Vasc Surg	R	Bologna, Italy	August–October 2016	31	31	Evaluation of CO2-DSA in arterial visualisation and graft deployment
Tantawy et al.26 (2021)	Ann Vasc Surg	R	Manchester, UK	2013–2019	15	9	Evaluation of post-operative renal function outcomes with CO2-DSA
Vacirca et al.27 (2022)	Ann Vasc Surg	R	Bologna, Italy	2016–2019	72	16	Evaluation of post-operative renal function outcomes with CO2-DSA and endoleak detection
Unal et al.29 (2023)	EJVES	R	Ankara, Turkey	2019–2022	34	29	Evaluation of post-operative renal function outcomes with CO2-DSA
Vacirca et al.28 (2023)	JVS	P	Multicentre†	2019–2021	65	19	Evaluation of CO2-DSA in arterial visualisation
Esposito et al.30 (2023)	Ann Vasc Surg	R	Florence, Italy	2019–2022	17	17	Evaluation of the feasibility and safety of a ‘zero contrast’ approach in patients with CKD
Quaglino et al31 (2024)	Ann Vasc Surg	R	Turin, Italy	2020–2021	52	52	Evaluation of post-operative renal function outcomes with CO2-DSA and endoleak detection
Chisci et al.11 (2025)‡	EJVES	P	Multicentre§	2023–2024	293	240	Technical results and insights of CO2-DSA in EVAR
Total					901	622

CKD = chronic kidney disease; CO₂ = carbon dioxide; DSA = digital subtraction angiography; EVAR = endovascular aortic repair; ICM = iodinated contrast medium; P = prospective; R = retrospective; RCT = randomised control trial; JVS: Journal of Vascular Surgery; Vasc and Endovasc Surg: Vascular and Endovascular surgery; EJEVS: European Journal of Vascular and Endovascular Surgery; Int J Cardiovasc Imaging: the International Journal of Cardiovascular Imaging; Ann Vasc Surg: Annals of Vascular Surgery.

^∗Comparative studies intended to compare CO₂-DSA with ICM alone.

^†Bologna, Italy; Malmö, Sweden; Münster, Germany; and Athens, Greece.

^‡Epub publication was in November 2024; printed in March 2025.

^§Florence (USL Toscana Centro), Turin, Padua, Verona, Florence (University of Florence), Perugia, Bologna, Genoa, Bari, and Brescia.

Nine studies^18,20,23, 24, 25^,28, 29, 30, 31 compared the results of standard ICM-EVAR with those of CO₂-EVAR.

Among the CO₂-EVAR group, 69.0% (622 patients) were performed with zero contrast medium and 31.0% were performed with adjunctive ICM (mean 20 ± 20 mL) (Supplementary Table S1).

The technical aspects of CO₂-DSA use are summarised in Table 2. Manual injection of CO₂ was performed in eight^18,19,21, 22, 23, 24^,26,32 studies published between 2007 and 2018. In all studies published after 2018,11, 25, 26, 27, 28, 29, 30, 31 an automatic injection system was adopted; specifically, the Angiodroid injector (Angiodroid S.p.A., San Lazzaro di Savena, Italy) was used in all cases. Mean CO₂ volume ejected to perform the procedure was 544 ± 408 mL.

Table 2.

Technical details of the procedures.

Author	CO2 patients – n∗	CO2 injection type	Mean CO2 – mL	Mean radiation dose – Gy			Mean fluoroscopy time – min			Duration of procedure – min
				CO2	ICM	p	CO2	ICM	p	CO2	ICM	p
Chao et al.18	16	Manual	–	925	529	.040	46	24	.010	180	138	.050
Lee et al.32	17	Manual	–	–	–	–	–	–	–	–	–	–
Huang et al.19	76	Manual	203	–	–	–	–	–	–	–	–	–
Criado et al.20	114	Angio Flush III Contrast Management System	395	–	–	–	21	28	.002	177	193	.010
Sueyoshi et al.21	40	Manual	–	–	–	–	–	–	–	–	–	–
De Angelis et al.22	13	Manual or automatic	–	–	–	–	–	–	–	–	–	–
De Almeida Mendes et al.23	16	Manual	156	–	–	–	31	30	.73	180	180	.55
Takeuchi et al.24	30	Manual	115	–	–	–	43	49	.57	172	171	.94
Mascoli et al.25	31	Automatic (Angiodroid)	–	–	–	–	20	–	–	170	–	–
Tantawy et al.26	15	Manual or automatic (Angiodroid)	–	182	–	–	35	–	–	–	–	–
Vacirca et al.27	72	Automatic (Angiodroid)	–	500	332	.001	–	–	–	–	–	–
Unal et al.29	65	Automatic (Angiodroid)	990	161	–	–	–	–	–	–	–	–
Vacirca et al.28	34	Automatic (Angiodroid)	–	–	–	–	16	20	.16	133	143	.10
Esposito et al.30	17	Automatic (Angiodroid)	946	–	–	–	22	–	–	–	–	–
Quaglino et al.31	52	Automatic (Angiodroid)	950	311	160	.001	15	13	.016	70	55	.006
Chisci et al.11,∗	293	Automatic (Angiodroid)	600	–	–	–	15	–	–	90	–	–
Total	901		544	416	340	–	26	27	–	147	147	–

CO₂ = carbon dioxide; ICM = iodinated contrast medium.

^∗Epub publication was in November 2024; printed in March 2025.

Post-operative renal function

Overall, seven studies compared PO-RFW after CO₂-EVAR with ICM-EVAR^18,20,23, 24, 25^,28, 29, 30, 31 (Table 3). Among the reported studies, one²⁹ reported a statistically higher rate of PO-RFW in the ICM-EVAR group than in the CO₂-EVAR group (p = .027); the other two studies showed no statistically significant difference.^24,31

Table 3.

Post-operative renal function worsening definitions and rates across the included studies.

Author	PO-RFW assessment method	Outcome metric or definition	CO2/ICM patients – n	Main results CO2vs. ICM	p value
Chao et al.18	sCr variation	Mean ΔsCr	16/84	CO2: ΔsCr 0.06 mg/dL ICM: ΔsCr 0.10 mg/dL	NS
Criado et al.20	eGFR variation	Mean % decrease in post-operative eGFR	114/22	ICM: 12% greater eGFR reduction compared with CO2	.040
De Almeida Mendes et al.23,∗	Mean sCr	Mean sCr ±SD	16/20	CO2: +1.1 ± 0.3 mg/dL ICM: +1.7 ± 0.5 mg/dL	.80
	Mean eGFR	Continuous (mean)	16/20	CO2: 79 mL/min/1.73 m2 ICM: 92 mL/min/1.73 m2	.22
Takeuchi et al.24	RIFLE classification	AKI stage (risk-injury-failure), n(%)	30/351	CO2: 1/30 (3.3%) ICM: 10/351 (2.8%)	.93
Vacirca et al.28	sCr variation	Mean sCr ±SD	72/249	CO2: +0.08 ± 0.04 mg/dL ICM +0.17 ± 0.09 mg/dL	.010
	eGFR variation	Mean ΔeGFR ±SD	72/249	CO2: −2.3 ± 1.1 mL/min/1.73 m2 ICM: −10.6 ± 5.3 mL/min/1.73 m2	.001
Unal et al.29,∗	AKI (threshold criteria)	≥25% rise or ≥0.5 mg/dL increase within 48 hours	34/34	CO2: 3% ICM:24%	.027
	Mean sCr	Mean sCr ±SD	34/34	CO2: 1.72 ± 0.59 mg/dL ICM: 1.75 ± 0.47 mg/dL	.29
	Mean eGFR	Continuous (mean)	34/34	CO2: 43 mL/min/1.73 m2 ICM: 39 mL/min/1.73 m2	.10
Quaglino et al.31	ARISE classification	AKI incidence, n(%)	52/49	AKI incidence 6% in both groups	NS
	Median sCr	Median sCr (IQR)	52/49	CO2: 1.1 mg/dL ICM: 0.98 mg/dL	.40

Δ = difference; AKI = acute kidney injury; ARISE = Aneurysm Renal Injury Score; CO₂ = carbon dioxide; eGFR = estimated glomerular filtration rate; ICM = iodinated contrast medium; IQR = interquartile range; NS = not statistically significant; PO-RFW = post-operative renal function worsening; RIFLE = Risk, Injury, Failure, Loss, End stage kidney disease; sCr = serum creatinine; SD = standard deviation.

^∗Matched cohorts of patients.

A study by Vacirca et al.²⁷ reported a statistically significant greater post-operative increase in sCr in the ICM-EVAR group than in the CO₂-EVAR group (p = .010).

A statistically greater reduction in eGFR was reported in two studies^20,27 after ICM-EVAR compared with CO₂-EVAR.

Secondary endpoints

Intra-operative technical details

Six^{18,20,23,24,29,31} of the included studies compared technical intra-operative data between CO₂-EVAR and ICM-EVAR (Table 2). Specifically, three studies^18,27,31 (two using automatic injection and one manual injection) reported a statistically higher radiation dose during the procedure in the CO₂-EVAR group than in the ICM group. Six studies^{18,20,23,24,29,31} analysed the total procedure fluoroscopy time, which was statistically longer in the CO₂-EVAR group in two of the included studies.^18,31 One study²⁰ reported a statistically longer fluoroscopy time in patients undergoing ICM-EVAR.

The mean duration of the CO₂-EVAR procedure was 147 ± 20 minutes. Two studies^18,31 described a statistically longer operation time for CO₂-EVAR, whereas one study²⁰ reported a shorter time.

Intra-operative imaging

Among the studies included, five^{24,26,27,29,32} analysed the accuracy of arterial visualisation to properly evaluate the proximal and distal landing zones during EVAR using CO₂ (Supplementary Table S2). Successful intra-operative visualisation of both renal arteries before graft deployment ranged between 53% and 100%. Successful visualisation of both hypogastric arteries to determine the distal landing zone occurred in between 94 and 100% of cases.^22,25,27 In the included studies, both the aortic bifurcation and the common iliac arteries were detected in all cases.

Table 4 shows the studies^21,23, 24, 25^,29,31 that evaluated the rate of intra-operative EL detection using CO₂ as the contrast medium. Overall, the findings indicated no statistically significant difference in EL detection between CO₂-EVAR and ICM-EVAR. When focusing on the detection of low flow type II ELs (T2ELs), two studies^25,31 reported a higher detection rate with CO₂-DSA; one study²¹ did not support this finding.

Table 4.

Endoleak detection.

Author	CO2 patients – n	EL evaluation	Synthesis of the results CO2vs. ICM
Chao et al.18	16	EL detection on final angiogram with both CO2 and ICM, one and six month imaging	No difference
Lee et al.32	17	EL detection on final angiogram with both CO2 and ICM	No difference
Huang et al.19	76	EL detection on final angiogram with both CO2 and ICM; comparison of sensitivity, specificity, PPV, and NPV for types I and II EL	(1)Lower sensitivity and PPV for CO2-DSA in type II EL detection. (2)Adequate for type I EL but less reliable than ICM
Sueyoshi et al.21	40	EL detection on final angiogram with both CO2 and ICM; CT follow up at six months	(1)Lower overall EL detection rate with CO2 (40%) vs. ICM (68%). (2)Same sensitivity (100%) for types I and III EL. (3)Lower sensitivity (50%) for type II EL with CO2 but better prediction of persistent type II EL.
Mascoli et al.25	31	EL detection on final angiogram with both CO2 and ICM	(1)Type II EL is more frequently detected with CO2 (10 vs. 2) (2)CO2 showed better agreement with CEUS (3)No difference in type I – III EL detection
Quaglino et al.31	52	EL detection on final angiogram	Better intra-operative type II EL detection with CO2 (25% vs. 14.3%, p = .20)

CEUS = contrast enhanced ultrasound; CO₂ = carbon dioxide; CT = computed tomography; DSA = digital subtraction angiography; EL = endoleak; ICM = iodinated contrast medium; NPV = negative predictive value; PPV = positive predictive value.

Risk of bias assessment

Four^21,22,30,32 of the included studies were deemed to be at high risk of bias, and 1211, 18, 19, 20, 24, 25, 26, 27, 28, 29, 31 were deemed to be at moderate risk of bias. None of the included studies were at low risk of bias.

DISCUSSION

This review represents a comprehensive analysis of the use of CO₂-DSA in the endovascular treatment of abdominal aortic aneurysms.

CO₂ is characterised by several beneficial characteristics due to its physical properties,¹² including low viscosity, high solubility, and rapid dissolution in blood, which allow for its safe injection into arteries below the diaphragm without causing significant gas embolism.^33,34 The solubility of CO₂ is 21 times greater than that of nitrogen, which facilitates its dissolution in the blood.^14,15 However, CO₂ injection above the diaphragm is contraindicated due to the potential for neurotoxic effects, such as disruption of the blood brain barrier, seizures, and loss of consciousness.³⁵

The literature shows an increased use of CO₂-DSA for peripheral and aortic endovascular procedures owing to its ability to deliver high quality imaging while minimising the risks associated with ICM.³⁶ Results on the effectiveness of CO₂-DSA in peripheral arterial disease¹⁰ are supported by strong evidence, and it is suggested as a valid alternative to ICM by the Global Vascular guidelines on chronic limb threatening ischaemia.³⁷ However, international guidelines for abdominal and iliac pathology¹³ do not yet provide specific recommendations on the use of this technique for EVAR.

Hence, the primary focus of this review was to assess the role of CO₂ in renal function protection.

Reporting of PO-RFW varied considerably across studies. While some used explicit definitions, others reported changes in sCr or eGFR. Due to this heterogeneity, a pooled analysis was not feasible, and outcomes were reported descriptively for each study. Overall, some studies reported promising results^20,27,29 on its role in renal protection.

Similarly, a study by Busutti et al.¹⁵ (not included in the review owing to potential patient overlap) confirmed that the use of CO₂ as a contrast medium, alone or combined with a low dose of ICM, was safer in terms of renal function than the standard dose of ICM in EVAR. Interestingly, the authors highlighted a potential role of ICM in the chronic kidney injury process, thus potentially impacting one year renal outcomes. This might define CO₂ as a valuable alternative to ICM for preventing PC-AKI and long term renal decline, not only in advanced CKD stages, where this relationship is more evident, but also in lower risk cases, considering its effect on patient survival after standard EVAR.³

Two different methods of CO₂-DSA injection can be used: manual and automatic. Manual injection, the historical standard, provides control over the injection process but involves higher radiation exposure and a greater risk of air contamination. In contrast, automated systems improve safety and reproducibility through precise control of volume and pressure.^9,28 Recent studies have adopted standardised protocols for pressure and frame rates, improving visualisation in both standard and complex aortic procedures.^9,12

In this review, manual injection was used in eight of 16 studies before 2018; since then, all studies have adopted automated systems. The reviewed data span from 2003 to 2022, reflecting the evolution of the technique. Notably, 69.0% of procedures were zero contrast. The mean ICM used in hybrid procedures was only 20 mL. Vacirca et al.²⁸ reported that 57% of patients received zero contrast procedures, and Chisci et al.¹¹ reported up to 82%, highlighting the growing feasibility of zero contrast procedures.^{19,21,25,30,31}

Arterial visualisation remains a crucial technical aspect of EVAR. Discrepancies were observed in the ability to visualise the lowest renal artery with CO₂-DSA, with reported success rates ranging from 53% to 100%. Owing to its buoyancy, CO₂ may provide suboptimal imaging when renal ostia originate from the posterior aortic wall or in large aneurysmal sacs.³⁹ Mascoli et al.²⁵ reported reduced visualisation in such cases, whereas Vacirca et al.²⁷ achieved better results with a femoral introducer instead of a pigtail catheter. Furthermore, Chisci et al.¹¹ demonstrated that adopting specific adjunctive manoeuvres, such as stepwise catheter repositioning and injection synchronisation, can markedly improve image quality, achieving an 82% success rate in zero contrast procedures.¹¹

The variability observed across studies may reflect differences in techniques, injection parameters, and operator experience, rather than intrinsic limitations of CO₂ angiography, which may explain why the 2024 European Society for Vascular Surgery guidelines did not yet endorse its routine use. Nevertheless, CO₂ and ICM should not be regarded as two competing imaging strategies but rather as complementary tools. When proximal or distal landing zone visualisation is incomplete with CO₂ alone, a small adjunctive dose of ICM can be safely used to refine anatomic definition without compromising the renal protective benefit of a primarily CO₂ guided protocol. This combined strategy represents a pragmatic balance between image quality and renal preservation.

Together with landing zone visualisation for graft implant, intra-operative EL detection remains a major point in pursuing EVAR technical success. Detection rates for T1ELs, T2ELs, T3ELs were similar between CO₂-DSA and ICM-DSA. On the other hand, results for T2EL detection are contrasting, with some studies reporting better visualisation with CO₂ than with ICM. Hence, further studies should be performed to clarify the role of CO₂ in EL detection.

CO₂ has also raised questions about its safety profile. A recent literature review on CO₂ related complications in EVAR³⁸ reported only minor intra- and peri-operative complications (4.9%), such as transient hypotension, nausea, or abdominal discomfort, all resolving within 24 hours and without post-operative sequelae. No major adverse events or 30 day deaths were attributed to CO₂. Radiation exposure remains a concern due to the need for higher frame rates (4–6 frames per second) during CO₂-DSA.³⁹ However, recent protocols use lower frame rates (two frames per second) and fusion imaging to mitigate exposure.⁴⁰ Hybrid rooms with advanced imaging and operator expertise further reduce angiography frequency and enhance image quality.

Thanks to growing clinical evidence, CO₂-DSA has been adopted increasingly in advanced endovascular aortic procedures, particularly for its potential to mitigate the risk of contrast induced nephropathy and preserve renal function.⁹ The expanding integration of CO₂-DSA with fusion imaging and advanced imaging techniques is paving the way for its routine intra-operative use. However, current evidence on its application in complex and advanced EVAR procedures prevents a comprehensive assessment at the present stage.

In thoracic endovascular procedures, CO₂ has also been advocated, not primarily as a contrast agent as used below the diaphragm, but as an alternative flushing medium to saline for thoracic endografts to displace air bubbles, thereby reducing the risk of air embolism during aortic arch repairs. While only a few studies³⁵ have been published so far, the preliminary data are promising and suggest the potential for broader validation and clinical adoption of this adjunctive technique.

Limitations

This review has several limitations. Most included studies were observational and non-randomised, with only one randomised controlled trial, resulting in limited evidence quality. Risk of bias was rated as moderate to high in all studies, particularly due to a lack of blinding, small sample sizes, and incomplete outcome reporting.

There was substantial heterogeneity in outcome reporting, particularly regarding definitions of PO-RFW, which were variably based on sCr, eGFR, or author defined criteria. This prevented pooled analyses and limited direct comparison between studies. A lack of consensus on renal function measures highlights the need for standardisation, with future studies encouraged to adopt consistent CO₂-DSA protocols and define renal function worsening, preferably using the Risk, Injury, Failure, Loss, End stage kidney disease classification.^5,8,17

Additionally, variations in CO₂ injection techniques, catheter types, imaging protocols, and operator experience may have influenced technical outcomes, such as arterial visualisation and radiation exposure. Specifically, suboptimal visualisation of the proximal and distal sealing zones, reported in some studies despite others achieving complete visualisation, remains a critical issue, as accurate assessment of these areas is crucial for procedural success. The use of CO₂-DSA in EVAR is still undergoing progressive refinements, and most published series reflect early or midterm learning curve experiences. As a result, the current evidence base is not yet mature enough to support formal recommendations in the latest guidelines;¹³ however, growing literature, standardisation of reporting outcomes, and comprehensive studies will fill this knowledge gap.

Conclusion

CO₂-DSA appears to be a promising alternative or adjunct to ICM in standard EVAR, with evidence suggesting a potential benefit in preserving renal function and reducing overall ICM volume. While visualisation performance and procedural feasibility are generally satisfactory, further studies using the latest available technologies and protocols are needed to confirm the current findings and support broader integration of CO₂-DSA into routine clinical practice.

Ethical statement

Not applicable, as this is a systematic review of published literature with no direct involvement of human or animal subjects.

Data availability

All data analysed in this study are derived from published literature. Extracted datasets are available from the corresponding author upon reasonable request.

Declaration of generative ai and ai-assisted technologies in the writing process

During the preparation of this work the authors used Chat-GPT to make minor edits to the text for grammar and spelling. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for it.

Conflicts of interest

M.G. and E.G. are consultants for Cook Medical for fenestrated and or branched endovascular aneurysm repair. The other authors declare that they have no competing interests.

Appendix A. Supplementary data

The following are the supplementary data to this article: