Comparison of procedural efficiency and safety between transradial and transfemoral approaches in elective diagnostic cerebral angiography: A retrospective cohort study

Abstract

Diagnostic cerebral angiography is essential for the treatment of cerebrovascular disease. The conventional transfemoral approach (TFA) carries risks of vascular complications and longer recovery. The transradial approach (TRA) is a promising alternative; however, more evidence, especially direct comparisons with TFA, is needed to support its use in diagnostic procedures. This single-center retrospective cohort study compared TRA and TFA in elective diagnostic cerebral angiography. Overall, 96 patients treated between February 2021 and January 2024 were included, and 1:1 propensity-score matching was used to create 48 matched patients per group based on arterial access. The primary endpoint was the rate of overall procedural complications within 24 hours. The secondary endpoints included puncture, procedure, fluoroscopy, and hemostasis time. Statistical analysis was performed using Student t test, Mann–Whitney U-test, chi-square test, and Fisher test. Both groups showed 100% success with arterial puncture and selective angiography. Total procedure time was shorter with TRA (44.15 ± 8.75 vs 48.23 ± 10.56 minutes; P = .04), and hemostasis time was faster (5.53 ± 2.13 vs 7.08 ± 3.27 hours; P <.01). Complications were significantly lower in the TRA group (2.1% vs 14.6%; P = .02), with access-site hematomas and 1 pseudoaneurysm occurring mainly in the TFA group. For diagnostic cerebral angiography, TRA showed similar success to TFA but with greater efficiency, faster hemostasis, and significantly fewer access-site complications. TRA may be considered a preferred access strategy for eligible patients; however, given the retrospective observational design, the findings should be interpreted with caution and confirmed in prospective randomized studies.

1. Introduction

Diagnostic cerebral angiography remains a cornerstone procedure in the evaluation and management of cerebrovascular disorders, offering unparalleled spatial and temporal resolution for visualizing intracranial and extracranial vasculature.^[1] Traditionally, the transfemoral approach (TFA) has been the most common arterial access route over the past few decades because the common femoral artery has a large caliber and accommodates a wide range of catheter systems. However, TFA has well-documented drawbacks, including the need for prolonged post-procedural bed rest to reduce access-site bleeding and a risk of local vascular complications such as hematoma, pseudoaneurysm, and retroperitoneal hemorrhage.^[2,3] These factors add to patient discomfort, prolong recovery, and increase healthcare resource utilization.

Over the last decade, the transradial approach has emerged as an attractive alternative that has dramatically changed practice paradigms, first in interventional cardiology and subsequently in neurointervention. A large body of evidence from cardiology indicates that transradial approach (TRA) significantly reduces major bleeding and vascular complications compared to TFA, while also enhancing patient comfort and facilitating early ambulation.^[4,5] In neurointervention, recent evidence – including a comprehensive systematic review/meta-analysis and evaluations of radial-specific catheter systems – supports the feasibility and safety of TRA, while also emphasizing access-specific challenges (e.g., spasm or other access-site events) that require appropriate technique and complication management.^[6-8] Encouraged by these benefits, neurointerventionalists have adopted TRA for both diagnostic and therapeutic procedures. A growing body of literature suggests that TRA for neuroangiography is feasible, safe, and may offer advantages in specific subsets of patients, such as obese patients, those with severe peripheral vascular disease, and those requiring early mobilization.^[9,6]

Despite this promising shift, the evidence base for TRA in purely diagnostic cerebral angiography calls for further strengthening. Many studies conducted on neurointervention have been retrospective, single-center series or have included heterogeneous patient populations undergoing both diagnostic and complex interventional procedures. These may confound analyses of procedural metrics and complication rates specific to angiography.^[6,7] Accordingly, we focused exclusively on elective diagnostic cerebral angiography to minimize confounding from device selection, larger-bore sheaths, and prolonged procedure times inherent to therapeutic interventions, thereby isolating access-related efficiency and complication outcomes.^[6,7] There is a need for contemporary, direct comparative studies that would concurrently evaluate a comprehensive set of outcomes, including but not limited to complication rates and procedural efficiency in detail, such as puncture time, total procedure duration, fluoroscopy exposure, and time to hemostasis. A clear, quantitative understanding of these metrics is critical for training and resource allocation and for informed clinical decision-making.

To address this gap in the literature, this single-center, retrospective cohort study was designed to directly compare transradial and transfemoral approaches for elective diagnostic cerebral angiography. We aimed to compare key procedural time metrics and the rate of peri-procedural complications between access strategies. We postulated that TRA would offer similar procedural success rates, shorter overall procedure and hemostasis times and lower access-site complication rates than conventional TFA.

2. Methods

2.1. Study design and setting

This retrospective, single-center, comparative cohort study was performed in the Department of Neurosurgery, Brain Science Center, Handan Hangang Hospital. The study protocol was approved by the Institutional Review Board of Handan Hangang Hospital. This study was based on the STROBE guidelines.

2.2. Participants

Patients who attended our institution between February 2021 and January 2024 for elective diagnostic cerebral angiography were screened for eligibility. Computerized patient records were searched for the relevant procedure codes.

The inclusion criteria were as follows: age 18 years or older; underwent comprehensive 4-vessel cerebral angiography; informed consent to the procedure obtained from the patient or guardian; skin intact at the site of intended puncture, with no infection, lesions, etc; Pre-procedure Doppler ultrasound evaluation confirming lack of significant thrombosis and stenosis within the targeted arterial segment; and full clinical and procedural information available.

Inclusion/exclusion criteria: severe thrombosis or stenosis of the planned access artery; presence of existing neurological dysfunction, which may cause confusion regarding procedural success; severe dysfunction of vital organs (heart, liver, kidneys); known allergy to iodinated contrast agents; and in patients planned for radial procedures, a positive Allen test, with failure of the palmar blush to return within 5 seconds after ulnar artery release, was an exclusion criterion in patients undergoing radial procedures.

2.2.1. Group allocation and sample size

Eligible patients were divided into 2 cohorts according to the arterial access route in the TFA and TRA groups. Access was chosen based on the patient’s preference after a detailed consultation with the treating physician prior to the procedure. Formal sample size calculation was performed a priori. Based on an anticipated reduction in the overall complication rate from 15% in the TFA group to 3% in the TRA group, at an alpha error of 0.05, with a power of 80%, the minimum requirement was 48 patients per group. Calculations were performed using Power software, version 3.1.

2.3. Methods and results

All procedures were performed by the same team of interventional neuroradiologists to minimize interoperator variability. For TFA, patients were positioned supine. After appropriate sterile preparation and draping, the common femoral artery was punctured under ultrasound guidance using a 19-gauge needle. This was followed by insertion of a 5-French sheath. Systemic heparinization (4000 IU) was administered intra-arterially. Aortic arch angiography was performed using a 5-French pigtail catheter and subsequent selective catheterization was performed using a 5-French Simmons-2 catheter. With regard to the TRA, when possible, the right radial artery was preferred. The patient’s wrist was hyperextended, and the artery was punctured using a 21-gauge micropuncture kit under ultrasound guidance. A 5-French radial sheath was inserted. A “radial cocktail” consisting of 200 μg of nitroglycerin and 4000 IU of heparin was administered via the inserted sheath. Arch and selective angiography were performed using a 5-French Jacky or Simmons catheter, respectively. Hemostasis was achieved with a dedicated radial compression device in the TRA group and a femoral closure device, or with manual compression in the TFA group.

Procedural success consisted of 2 elements

1. Successful puncture: Stable sheath placement confirmed by pulsatile blood return.

2. Successful angiography determination of diagnostic-quality images (defined as complete opacification permitting definitive interpretation) of all 4 target vessels (both internal carotid and vertebral arteries). Angiographic quality was independently assessed by 2 neurointerventionalists. Disagreements were resolved through consensus. Procedures that met both criteria were considered successful.

Primary Outcome: The rate of procedural complications, in general, was defined as the development of any of the following within the first 24 hours post-procedure: access-site hematoma, as evidenced by clinically significant swelling or ecchymosis requiring prolonged compression or intervention; symptomatic vasospasm requiring pharmacological treatment; and pseudoaneurysm confirmed by duplex ultrasonography.^[4,10]

Secondary Outcomes: Procedural metrics included: puncture time (from local anesthesia infiltration to successful guidewire insertion), total procedure time (from skin preparation to sheath removal), fluoroscopy time (total fluoroscopy time recorded by the angiographic system), and time to hemostasis (from sheath removal to achievement of effective hemostasis at the puncture site, after which the compression device was secured). The success rates of arterial and selective angiography were also recorded.

Outcome assessors (for example, nurses or physicians recording any complications after the procedure) were not aware of the tested hypothesis; complications were adjudicated by a blinded independent neurologist based on predefined criteria.

2.4. Propensity-score matching

Owing to the retrospective nonrandomized study design, patients were stratified to either TRA or TFA based on physician preference, which could be a source of selection bias. Furthermore, to reduce the effect of confounding factors on procedural outcomes, patients who underwent TRA or TFA were matched for propensity scores.

These included age, sex, body mass index, history of hypertension, diabetes mellitus, active smoking, previous cerebrovascular events, and current use of antiplatelet or anticoagulant agents.^[11] Antiplatelet therapy was defined as the use of aspirin and/or P2Y12 inhibitors, whereas anticoagulant therapy was defined as the use of vitamin K antagonists or direct oral anticoagulants; these variables were recorded and presented separately in the baseline characteristics. The propensity score for TRA (vs TFA) was calculated using a logistic regression. Matching was carried out 1:1 with a caliper width of 0.02 using nearest-neighbor matching. Propensity-score matching was performed before any between-group outcome comparisons. All primary and secondary outcomes were subsequently analyzed in the matched cohort, and no additional regression adjustment was applied post-matching. Post-matching balance was assessed using standardized mean differences (SMDs) and variance ratios for each covariate, and the distribution of propensity scores between groups was compared numerically. An SMD of <10% was considered indicative of adequate covariate balance after matching.

2.5. Statistical analysis

The analyses were performed using IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp., Armonk). The normality of continuous variables was tested using the Shapiro–Wilk test. Normally distributed data are presented as mean ± standard deviation and were compared between groups using the independent samples Student t test. Skewed data are reported as median (interquartile range) and were compared using the Mann–Whitney U-test. Categorical variables are presented as numbers (percentages), and differences were analyzed using the chi-square test or Fisher exact test, as appropriate. In the matched cohort, effect estimates are additionally reported with 95% confidence intervals (mean difference for continuous outcomes; risk difference and odds ratio for binary outcomes). A 2-sided P-value of <.05 was regarded as statistically significant. No imputation was performed for missing data, as only patients with complete records were included.

2.6. Ethical and compliance statements

This study was approved by the Institutional Review Board of Handan Hangang Hospital (approval no. HGYY-KYLX-2022-03) and was conducted in accordance with the ethical standards of the institutional committee and the Declaration of Helsinki and its later amendments. Written informed consent was obtained from all participants for both the procedure and use of anonymized data for research purposes. All patient identifiers were removed to protect privacy. As this was a retrospective study, it was exempt from clinical trial registration. The authors declare that they have no conflicts of interest.

2.7. Data availability

Due to the sensitive nature of the clinical data and patient privacy protections, the raw datasets generated during this study are not publicly available. De-identified data can be made available to qualified researchers upon reasonable request to the corresponding author, subject to a formal data sharing agreement and approval by the local ethics committee.

3. Results

3.1. Procedural metrics

After matching, a SMD of <10% for all covariates was suggestive of good matching between the groups (Table 1). All subsequent analyses were performed on the matched group (48 patients in each group, n = 96). After matching, all baseline covariates demonstrated SMDs of <10%, indicating adequate balance between the TRA and TFA groups.

Table 1.

Baseline characteristics of the matched cohorts.

Variable	TRA group (n = 48)	TFA group (n = 48)	SMD
Age, years	58.4 ± 12.3	59.1 ± 11.8	0.06
Male sex, n (%)	28 (58.3)	26 (54.2)	0.08
BMI, kg/m2	25.1 ± 3.4	24.8 ± 3.6	0.08
Hypertension, n (%)	32 (66.7)	30 (62.5)	0.09
Diabetes mellitus, n (%)	15 (31.3)	14 (29.2)	0.04
Smoking history, n (%)	18 (37.5)	16 (33.3)	0.09
Prior cerebrovascular disease, n (%)	10 (20.8)	11 (22.9)	0.05
Pre-procedural antiplatelet/anticoagulant use, n (%)	12 (25.0)	10 (20.8)	0.1

Data are presented as the mean ± standard deviation or number (percentage). SMD <0.10 indicate adequate balance between groups after matching.

SMD = standardized mean difference.

Patients with incomplete clinical or procedural information were excluded at screening per the prespecified inclusion criteria; no additional exclusions occurred after eligibility determination and propensity-score matching. Ninety-six patients were included in the final analysis, with 48 in each group. Table 2 summarizes the comparison of the key procedural time metrics between the TRA and TFA groups, including mean differences (TRA–TFA) with 95% confidence intervals. The total procedure time was significantly shorter in the TRA group (44.15 ± 8.75 minutes) compared to the TFA group (48.23 ± 10.56 minutes; mean difference, −4.08 minutes; 95% CI, −8.01 to − 0.15; P = .04). Similarly, time to hemostasis (5.53 ± 2.13 vs 7.08 ± 3.27 hours; mean difference, −1.55 hours; 95% CI, −2.67 to − 0.43; P <.01) were significantly reduced in the TRA group. In contrast, fluoroscopy time was slightly longer in the TRA group (8.27 ± 6.18 minutes) than in the TFA group (7.04 ± 6.64 minutes; mean difference, 1.23 minutes; 95% CI, −1.37 to 3.83), but this difference was not statistically significant (P = .35). The puncture time was slightly longer in the TRA group (4.81 ± 3.93 minutes) than in the TFA group (3.51 ± 3.23 minutes; mean difference, 1.30 minutes; 95% CI, −0.16 to 2.76), but this difference was not statistically significant (P = .08).

Table 2.

Comparison of procedural time metrics between groups.

Variable	TFA group (n = 48) mean ± SD	TRA group (n = 48) mean ± SD	Mean difference (TRA–TFA)	95% CI	P
Puncture time (min)	3.51 ± 3.23	4.81 ± 3.93	1.3	−0.16 to 2.76	.08
Total procedure time (min)	48.23 ± 10.56	44.15 ± 8.75	−4.08	−8.01 to − 0.15	.04
Fluoroscopy time (min)	7.04 ± 6.64	8.27 ± 6.18	1.23	−1.37 to 3.83	.35
Hemostasis time (h)	7.08 ± 3.27	5.53 ± 2.13	−1.55	−2.67 to − 0.43	<.01

Mean differences are presented as TRA minus TFA with corresponding 95% confidence intervals.

TFA = transfemoral approach, TRA = transradial approach.

3.2. Procedural success rates

Arterial puncture and selective 4-vessel cerebral angiography were successful in all 96 patients; thus, both endpoints were reached in each group with a 100% success rate (Table 3). Given identical success rates, no between-group statistical comparison was performed for these endpoints.

Table 3.

Comparison of procedural success rates between groups.

	n	Puncture success rate (%)	Arteriography success rate (%)
TFA group	48	100	100
TRA group	48	100	100
P	–	–	–

Both groups achieved a 100% procedural success rate. Therefore, no statistical comparison was performed.

TFA = transfemoral approach, TRA = transradial approach.

3.3. Complication rates

The overall complication rate was significantly lower in the TRA group (2.1% [1/48]) than that in the TFA group (14.6% [7/48]; P = .02; Table 4), corresponding to an absolute risk reduction of 12.5% (risk difference, −12.5%; 95% CI, −26.8% to 3.7%) and a number needed to treat of 8 to prevent 1 complication within 24 hours; the relative risk was 0.20 (95% CI, 0.04 to 1.10). All complications in the TFA group were access-site related: 5 patients (10.4%) developed a clinically significant hematoma and 1 patient (2.1%) developed a pseudoaneurysm, which was confirmed by duplex ultrasonography. One patient (2.1%) in the TFA group developed symptomatic vasospasm that required pharmacological treatment. In the TRA group, 1 complication (2.1%) was clinically significant access-site hematoma. No cases of symptomatic vasospasm or pseudoaneurysms were observed in the TRA group. The hematoma rate was significantly different between groups (P = .02).

Table 4.

Post-procedural complications within 24 hours.

Complication	TFA group (n = 48) n (%)	TRA group (n = 48) n (%)	Risk difference	95% CI	Odds ratio	95% CI	P
Hematoma	5 (10.4)	1 (2.1)	−8.3%	–	–	–	.02
Pseudoaneurysm	1 (2.1)	0 (0)	−2.1%	–	–	–	1
Symptomatic vasospasm	1 (2.1)	0 (0)	−2.1%	–	–	–	1
All complications	7 (14.6)	1 (2.1)	−12.5%	−23.3% to − 1.7%	0.12	0.01–1.06	.02

Effect estimates are reported as risk difference and odds ratio with 95% confidence intervals.

TFA = transfemoral approach, TRA = transradial approach.

4. Discussion

Procedures were evaluated for technical success rate, total procedure time, hemostasis time, and total complications occurring within 24 hours for both TRA and TFA in this collection of cases. This observational study demonstrated that, with a 100% technical success rate, TRA was associated with shorter total procedure and hemostasis times, and a lower 24-hour complication rate compared with TFA.

The significantly lower total procedure time in the TRA group (44.15 ± 8.75 minutes) compared with that of the TFA group (48.23 ± 10.56 minutes; P = .04) is in keeping with the expanding literature that recommends TRA in neurointerventions. This could be attributed to the fact that there is no need for prolonged bed stay after the procedure, making patient mobilization easier and earlier discharge for planning.^[12] Conversely, the slightly longer puncture time in the TRA procedure (4.81 vs 3.51 minutes), though not statistically significant (P = .08), may reflect the recognized learning-curve for radial artery cannulation despite the use of a standardized protocol by an experienced operator team.^[13] Additionally, the slightly higher fluoroscopy time for TRA, although not significant, could be attributed to the relative difficulty of navigating the catheter from the radial artery to the aortic arch, which is more complicated than that of the femoral route. Although the difference in total procedure time was statistically significant, the absolute reduction was modest (approximately 4 minutes) and may have limited clinical impact for an individual elective diagnostic case. Nevertheless, even small per-case time savings may translate into meaningful cumulative gains in workflow efficiency and throughput in high-volume angiography units, particularly when coupled with the shorter time to hemostasis observed with TRA. Conversely, the numerically higher fluoroscopy time with TRA – despite not reaching statistical significance – highlights the need to monitor radiation-related metrics and reinforce radiation-minimizing techniques during TRA adoption (e.g., optimizing catheter selection, minimizing unnecessary runs, and addressing navigation challenges related to arch anatomy and the radial learning-curve).

However, the overall procedure duration was not prolonged. The most remarkable difference was in the time to hemostasis: 5.53 ± 2.13 hours for TRA versus 7.08 ± 3.27 hours for TFA; P <.01. This reduction is the critical practical benefit of TRA because of the superficial position of the radial artery that allows for efficient compression with minimal discomfort to the patient and for the immediate ambulation of the majority of patients, unlike the very prolonged immobilization required following femoral access.^[14]

A key finding was that the overall complication rate was lower in the TRA group than in the TFA group (2% vs 14.6%, P = .02). All complications seen in the TFA group were related to access sites, mainly hematomas (10.5%) and 1 pseudoaneurysm (2%), while the single complication in the TRA group was hematoma (2%). This finding suggests a potentially favorable safety profile of the radial approach to vascular access in neuroangiography, particularly with respect to access-site vascular complications. The fact that no symptomatic vasospasm occurred in the TRA group is intriguing since spasm of the radial artery is one of the most frequently mentioned drawbacks. Intra-arterial administration of nitroglycerin according to our protocol likely contributed to this favorable outcome, which is also supported by previous studies.^[8] Taken together, these findings suggest that TRA may reduce access-site injury in selected patients; however, causal inference is limited by the observational design.

A theoretical implication of this study is the support of the TRA as an established fallback plan for diagnostic neuroangiography among selected patients, countering the classic femoral-first approach. It helps establish a risk-benefit ratio where the trade-off for a slightly more difficult access is a substantial improvement in patient satisfaction, security, and efficiency. A practical application of these findings is evident in clinical practice and resource allocation in hospitals. First, practitioners may consider TRA as an initial approach in appropriately selected patients, particularly in those with a higher bleeding risk or those requiring early ambulation. Second, the shorter hemostasis time could potentially result in enhanced catheterization laboratory efficiency and reduced nursing workload for monitoring patients after the procedure for vascular access complications.^[7] Hospital policies should provide educational support to overcome the learning-curve of transradial neurointervention.

Several limitations of this study are intrinsically related to its retrospective, single-center nature. First, patient selection for either TRA or TFA was performed based on consultant judgment/consumer preference following discussion. This nonrandom allocation of patients to the procedure type is a substantial generator of the risk for selection bias. For example, patients who were younger, more active, or had occupations that necessitated rapid ambulation would potentially be more likely to opt for a TRA. Conversely, proceeding with the more established TFA technique would be operator preference for cases that were likely to pose anatomical issues (e.g., patients with poor radial pulse or significant tortuosity) or for their personal experience with the learning-curve of the radial approach. While no significant differences were seen in major areas of patient selection for both groups with regard to major bases (Table 1), there could be certain factors of residual confounding that could influence both procedure selection as well as both procedure endpoints of procedural time and complications. These factors could have remained either unknown or unaccounted for. Importantly, propensity score matching can only address measured covariates and cannot eliminate confounding from unmeasured or imperfectly measured factors. Potential residual confounders include vascular anatomy (e.g., aortic arch type, supra-aortic tortuosity, peripheral arterial disease burden), access-vessel caliber and spasm propensity, antithrombotic intensity beyond “use” status, and peri-procedural factors such as sedation strategy or hemostasis management. In addition, operator and institutional experience with TRA may have evolved over time (learning-curve and temporal effects), which could influence procedure time, fluoroscopy usage, and complication ascertainment despite matching. Therefore, the observed associations should be interpreted as comparative effectiveness in this matched cohort rather than unbiased estimates of causal efficacy. Indeed, the extent to which the findings of this study can be generalized is dubious and potentially restricted to this group of patients. Second, the study group was of sufficient number to test for differences in the rates of primary complications, but potentially too small to test for rare adverse events or differences in certain procedural measures. Third, it is likely that blinding was not achieved for the assessment of the access-site, given its subjective nature, potentially causing observational bias despite the objective selection of complications. Lastly, complications could not be evaluated past the 24-hour point, potentially including delayed complications, such as pseudoaneurysm or radial artery occlusion.

Future studies should attempt to solve these issues by conducting prospective, randomized, multi-center trials with more study subjects for an extended follow-up duration to ensure the long-term safety and generalizability of their findings. Additionally, studies should be conducted to determine the cost-effectiveness of TRA over TFA, taking into consideration the variations in the management of complications with respect to the length of stay. Moreover, analysis of patient preference scores between the 2 techniques would also help obtain more information about patient experience.

Acknowledgments

The authors would like to thank the Department of Neurosurgery, Brain Science Center, Handan Hangang Hospital for their support and contributions to this study.

Author contributions

Conceptualization: Guangtao Xu.

Data curation: Guangtao Xu.

Formal analysis: Guangtao Xu.

Investigation: Manyi Zhao.

Methodology: Manyi Zhao.

Resources: Yabo Kong, Jiangying Wang, Rongrong Wang.

Software: Jiangying Wang.

Supervision: Guangtao Xu.

Validation: Yabo Kong.

Writing – original draft: Guangtao Xu.

Writing – review & editing: Rongrong Wang.