Article Text

Download PDFPDF

Original research
Maturing institutional experience with the transradial approach for diagnostic cerebral arteriography: overcoming the learning curve
  1. Benjamin M Zussman1,
  2. Daniel A Tonetti1,
  3. Jeremy Stone1,
  4. Merritt Brown2,
  5. Shashvat M Desai2,
  6. Bradley A Gross1,
  7. Ashutosh Jadhav2,
  8. Tudor G Jovin3,
  9. Brian Thomas Jankowitz1
  1. 1 Neurosurgery, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
  2. 2 Neurology, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
  3. 3 Neurology, Cooper University, Camden, NJ, USA
  1. Correspondence to Dr. Brian Thomas Jankowitz, Department of Neurological Surgery, University of Pittsburgh, Pittsburgh, PA 15213, USA ; ankbt{at}


Background Despite growing interest in the transradial approach for neurovascular procedures, prospective data about the learning curve for neurointerventionalists adopting this approach are limited.

Methods A subsequent prospective series of 50 consecutive right transradial diagnostic cerebral arteriograms was compared with our initial institutional experience using a procedural staging system. The primary outcome was the ability to achieve the predefined procedural goals using the radial approach. Secondary outcomes included the technical ability to access and inject each supraaortic artery of interest and the incidence of complications.

Results The primary outcome was achieved in 49 patients (98%) compared with 88% in the initial series (p=0.05). One stage 2 failure (2%) occurred. Crossover to the transfemoral approach occurred in one patient (2%) compared with 8% in the initial series (p=0.16). All supraaortic arteries of interest were accessed and injected with success rates between 93% and 100%. There were no major complications and two minor complications.

Conclusion Neurointerventionalists can overcome the right transradial learning curve and achieve high success rates and low crossover rates after performing 30–50 cases.

  • transradial
  • radial artery
  • diagnostic cerebral angiography technique
  • learning curve

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The field of interventional cardiology has adopted a transradial-first approach over the traditional transfemoral approach because it is associated with lower bleeding and vascular complications and improved patient safety.1 These clinical benefits have generated interest in adopting the radial approach for neuroendovascular procedures,2 but it is unknown if the radial approach is associated with clinical advantages over the femoral approach in this domain as well. Before a randomized comparison between the radial and femoral approaches can be meaningfully conducted, neurointerventionalists must overcome the radial learning curve and achieve technical proficiency.

Interventional cardiology studies suggest that the threshold to overcome the basic learning curve is approximately 30–50 transradial cardiac catheterizations.3–5 Thereafter, procedural success rates remain high and complication rates remain low, although operators do become increasingly efficient with additional experience.6

In a prospective study of 50 consecutive patients at our institution,7 we recently showed that neurointerventionalists transitioning to the radial approach achieve moderate early success and a low complication rate. However, prospective data regarding the learning curve for transradial cerebrovascular procedures are lacking. The aim of this study was to prospectively study a subsequent 50 consecutive patients to characterize the relationship between experience and performance.


Trial design

The study was a prospective, single-arm, single-center registry of 50 consecutive eligible patients who underwent right transradial diagnostic cerebral arteriography at a large academic medical center in the USA. The study protocol was approved as a quality improvement project by the institutional Quality Improvement Review Committee. The study was designed and conducted by the academic investigators. Funding was provided by MicroVention (Aliso Viejo, California, USA) and the Pittsburgh Foundation (Pittsburgh, Pennsylvania, USA); the funding was used for purchasing equipment necessary for transradial angiography.

Patient selection

All inpatients and outpatients aged 18 years or older were eligible for inclusion in the study if they were referred for diagnostic cerebral arteriography. The goal(s) of the procedure were explicitly defined (table 1) prior to the procedure. Patients were excluded if they had relative contraindications to transradial arteriography that have been previously defined.7 The study was concluded after 50 patients underwent attempted transradial diagnostic cerebral angiography.

Table 1

Diagnostic goals of the procedures performed

Operator experience

We previously reported our initial institutional experience with adopting the right transradial approach for diagnostic cerebral arteriography in a prospective consecutive series of 50 patients.7 At the conclusion of that study, our cumulative contemporary institutional transradial case volume was 91 cases, including diagnostic and interventional cerebrovascular procedures. Thereafter, we continued to perform radial cerebrovascular procedures as a regular practice for 3 months until the beginning of the current study. At the beginning of the current study our cumulative institutional transradial case volume was 146 cases, including diagnostic and interventional procedures. Each neurointerventional operator had experience independently performing more than 100 transfemoral diagnostic cerebral arteriograms and at least 18 right transradial cerebral arteriograms (range 18–74) prior to the study.

Procedural protocol

Transradial arteriography was performed with a standardized procedural workflow that has been previously described.7 Diagnostic cerebral arteriography was performed using a variety of diagnostic catheters including 100 cm hydrophilic-coated 5F Simmons 1, Simmons 2, and also Vert catheters at the discretion of the neurointerventionalist. The ability to access and inject each supraaortic artery of interest was recorded. An arch aortogram was not routinely performed.

At the conclusion of the procedure an inflatable transradial tourniquet was applied, the sheath was removed, and the hemostatic band was inflated to achieve patent hemostasis. The band was deflated per institutional protocol over 1 hour. The presence or absence of a right radial artery palpable pulse was evaluated by the investigators and recorded prior to discharge for outpatients or within the first 24 hours post-procedure for inpatients.

Primary outcome

The primary outcome was the ability to achieve the predefined goals of diagnostic cerebral arteriography (table 1) using the transradial approach. A previously described procedural staging system (box 1) was used to further subclassify the procedure into formalized stages 1–3b. The results from this series of patients were then compared with the previously reported initial series of 50 patients.7

Box 1

Procedural staging system for transradial diagnostic cerebral arteriography

  • Stage 1: the right radial artery is punctured and a sheath is inserted into the artery

  • Stage 2: a diagnostic catheter is advanced to the aortic arch

  • Stage 3a: the goals of arteriography are achieved with standard techniques, meaning the use of traditional 4F or 5F catheters, 0.035 inch or 0.038 inch guidewires, and exchange-length J-wires

  • Stage 3b: the goals of arteriography are achieved with adjunctive techniques, meaning any additional non-diagnostic catheters or wires such as long sheaths, intermediate catheters, microcatheters, and microwires

Crossover to the transfemoral approach meant the decision was made to convert the access site from the radial artery to the femoral artery. Termination of the procedure meant the decision was made to terminate the procedure because the risk or cost of proceeding with further arteriography outweighed the potential benefits of proceeding. Either crossover or termination could occur if the goals of arteriography could not be achieved using the transradial approach.

Secondary outcomes

Secondary outcomes included the ability to access and inject the supraaortic arteries using the transradial approach. Supraaortic arteries were deliberately selected for catheterization based on the predefined goals of each procedure and injection of arteries that would not directly fulfill the goals of the procedure was avoided. ‘Accessed and injected’ meant the catheter was positioned securely within the artery of interest and an arteriogram was performed. ‘Attempted but unable’ meant the catheter could not be positioned securely within the artery of interest.

Other secondary outcomes were designed to evaluate safety of the transradial approach. Major complications were defined as radial artery occlusion, critical hand ischemia, compartment syndrome, arteriovenous fistula development, radial artery pseudoaneurysm, radial artery avulsion or eversion, radial nerve injury, stroke, intracranial hemorrhage, or death. Minor complications were defined as radial artery vasospasm, local forearm ecchymosis, and local forearm hematoma. Procedure-related complications were recorded by the operator at the immediate conclusion of the procedure and during post-procedure evaluation of the radial artery pulse. Fluoroscopy time was recorded at the immediate conclusion of each procedure.

Statistical analysis

The results from this study were compared with the results from our initial set of 50 patients.7 Between-group comparisons for categorical variables were performed using χ2 or Fisher’s exact test, as appropriate. For continuous variables, means between groups were compared using the t-test. SPSS software version 23.0 (IBM, Armonk, New York, USA) was used.


Patient cohort

Beginning in January 2019 and for 34 total days, 57 consecutive patients referred for diagnostic cerebral arteriography without the possibility of intervention were screened for inclusion in the trial. Seven patients (12%) were excluded; two had a history of right transradial arterial line or arteriogram within the past 30 days, one had a recent right upper extremity trauma, one had a right upper extremity arteriovenous fistula for hemodialysis, one had no palpable right radial artery pulse, one had congenital agenesis of the right upper extremity, and one patient had left hemiplegia and would not be able to adhere to postoperative avoidance of high-resistance wrist movements with transfers. The remaining 50 patients constituted the transradial study cohort and their baseline characteristics are shown in table 2.

Table 2

Patient characteristics

Procedural details

In three cases (6%) the first diagnostic catheter was exchanged for a second diagnostic catheter to complete the procedure, and one case (2%) required a third diagnostic catheter. The remaining 46 cases (92%) were completed using only a single diagnostic catheter.

Primary outcome

The predefined goals of diagnostic cerebral arteriography were achieved using the transradial approach in 49 patients (98%), constituting the primary outcome (table 3). There were no stage 1 failures. Stage 2 failure occurred in one patient (2%); a radioulnar loop was encountered, the diagnostic catheter could not be advanced to the aortic arch and crossover occurred. There were no stage 3a or stage 3b failures. For comparison, the results of our initial series are also shown.

Table 3

Primary outcome

Secondary outcomes

Injection success rates for all supraaortic arteries were between 93% and 100%. Artery-specific results are listed in table 4; for comparison, the results of our initial series are also shown. The rate of successful access and injection of the left vertebral artery significantly improved between cohorts (59% vs 100%, p=0.014).

Table 4

Interval improvement in technical success by supraaortic artery

There were no major complications in this study. All 50 patients were examined following the procedure and all (100%) had a palpable right radial artery pulse. There were two cases (4%) of minor intraprocedural vasospasm, both of whom responded favorably to additional vasodilators and subsequently had successful completion of the study.

Fluoroscopy time was recorded and compared with the initial series (unpublished data). In the current study there were significant decreases in fluoroscopy time (15.7 vs 22.8 min, p=0.002) despite a similar number of vessels accessed and injected (293 vs 285 vessels injected). The fluoroscopy time per vessel accessed and injected for this study was 3.4±3.9 min.


As operators develop experience with an alternative surgical approach, technical performance typically improves rapidly and then plateaus as the basic learning curve is mastered. In interventional cardiology, the radial learning curve has been defined by comparing radial learners with radial experts who have completed hundreds of radial cases (typically >300). Ball et al compared radial learners with radial experts and found that learners needed to perform 50 transradial coronary interventions before their clinical outcomes were not inferior to the outcomes of the radial experts.3 Similarly, a study of the National Cardiovascular Data Registry showed that markers of operator performance like procedural success and radiation exposure began to plateau once operators performed 30–50 transradial coronary interventions.4

Snelling et al showed that individual neurointerventionalists learning the radial approach rapidly improve their overall efficiency between case numbers 1–5 and case numbers 11–15.8 Aside from this retrospective data about early learning, the right radial learning curve for neurointerventionalists has not previously been defined.

The aim of this study was to prospectively characterize the experience–performance curve for operators learning the right radial approach for diagnostic cerebral arteriography. At the beginning of this study our team of operators had been routinely employing the radial approach for nearly 4 months and operators had independently performed between 18 and 74 radial cases. With this institutional experience, the predefined goals of angiography were achieved using the radial approach 98% of the time. From our initial study7 to this study, the rate of stage 1 failure decreased from 6% to 0% (p=0.02), the rate of stage 3 failure decreased from 4% to 0% (p=0.11), and the transfemoral crossover rate decreased from 8% to 2% (p=0.16).

The standardized procedural workflow did not change between our initial series and this study and we identify several factors that we think contributed to our improved performance. Ultrasound-guided radial artery cannulation can be frustrating at first but becomes routine with practice. Reformatting Simmons-shaped catheters in the aortic arch can be time consuming in the beginning but is ultimately a simple skill that can be mastered with repetition; Lee et al 9 and Snelling et al 8 have described practical strategies to perform this maneuver. Catheter manipulation and steering into the supraaortic vessels can be awkward and inefficient early on but rapidly becomes intuitive as operators mature in their understanding of this different approach to the aortic arch.

Procedural efficiency improved between the two cohorts, mirroring the advancement in technical ability with practice. From the initial study to this study, there was significantly less fluoroscopy time used (22.8 vs 15.7 min, p=0.002), despite a similar number of vessels injected. This decrease in fluoroscopy time is largely due to improved operator efficiency navigating the aortic arch from the radial approach.

There was one stage 2 failure (2%) in this study due to an anatomical variant that could not be overcome. Vascular anatomical variations of the upper extremity are common and are typically manageable abnormalities, but they do rarely lead to access site crossover.10 Radial angiograms were obtained after sheath access in every case in this study, which aided in the identification and management of radial artery variants. Although the incidence of radioulnar loops in the general population is approximately 1%, there were four patients (8%) in this study with true radioulnar loops. We were able to overcome this anatomical variant in three of the four cases (75%) to successfully complete the procedure, contributing to our low crossover rate in the second cohort.

Although the right radial approach for cerebrovascular procedures has technical limitations, increasing experience led to interval improvement in the ability to access and inject supraaortic arteries. From the initial study to this study, the ability to access and inject the left vertebral artery improved from 59% to 100% (p=0.014). In both studies access of the left vertebral artery was only attempted when injection of that vessel was necessary to directly fulfill the predefined diagnostic goals of the procedure. At present, our strategy for evaluating the posterior circulation from the right radial approach is as follows: if the right vertebral artery is dominant or co-dominant, we firmly hand inject or power inject (typically 8 mL/s, 12 mL volume) the right vertebral artery and are often able to adequately evaluate the contralateral posterior circulation from that injection alone; if a firm right vertebral artery injection is not adequate to evaluate the contralateral posterior circulation, then we attempt to access the left subclavian or vertebral artery, as necessary; if the right vertebral artery is diminutive or otherwise inadequate, then we attempt to access the left subclavian or vertebral artery. Fortunately, in cases when the right vertebral artery is diminutive, the left vertebral artery is commonly robust and therefore easier to access and inject. It is often useful to use Simmons shapes 2 and 3 catheters, which have longer tails, to catheterize the left vertebral artery when the left subclavian artery is elongated.

When teams work together to develop experience with an alternative surgical approach, the learning curve of individual operators may be accelerated dramatically because colleagues rapidly disseminate best practices and strategies to avoid pitfalls to the entire group.11 Developing familiarity with an alternative approach among non-operator team members, such as technologists and nurses, is a gradual parallel process that impacts room set-up, patient positioning, and effective intraprocedural assistance. Because individual operator learning and group learning are interrelated processes, it is difficult to define the case volumes necessary to achieve competency with the radial approach in isolation; our institutional experience suggests that a busy neurointerventional practice can overcome the basic radial learning curve in several months and after operators each perform about 30–50 cases.

Study limitations

The experience–performance curve described in this study may not be generalizable to centers with one or few neurointerventionalists who may not benefit from the group learning effect. Similarly, centers with low procedural volumes may take longer to achieve competence with the radial approach. The learning curve for diagnostic cerebral arteriography may differ from the learning curve for transradial cerebrovascular interventions.

It is difficult to know if our institutional learning curve has fully plateaued despite the 98% procedural success rate. Indeed, interventional cardiology studies suggest that, even once the basic radial learning curve plateaus, operator proficiency continues to improve with additional experience.6 Our results suggest that right transradial diagnostic cerebral arteriography can be performed with a high level of efficacy and safety, and prospective comparison with the transfemoral approach should be the focus of future inquiry.


Neurointerventionalists can overcome the right transradial learning curve and achieve high success rates and low crossover rates for cerebral arteriography after performing 30–50 diagnostic cases.



  • Contributors Drafting the article: BMZ, DAT. Study conception: BMZ, DAT, BTJ. Data acquisition: BMZ, DAT, JS, MB. Data analysis: SMD, BMZ, DAT, BTJ. Study supervision: AJ, TGJ, BAG, BTJ. Critical revision of manuscript: All authors. Final approval of manuscript: All authors.

  • Funding This work was supported by grant number UN2018-ARMPITT from the Pittsburgh Foundation and by grant entitled ’Arterial Radial Management at UPMC' from Microvention Inc.

  • Competing interests BTJ: Consultant: Medtronic. TGJ: Consultant: Stryker Neurovascular; Ownership Interest: Anaconda; Advisory Board/Investor: FreeOx Biotech; Advisory Board/Investor: Route92; Advisory Board/Investor: Blockade Medical; Consultant; Honoraria: Cerenovus. BAG: Consultant: Microvention.

  • Ethics approval This study was approved by our institutional Quality Improvement Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Patient consent for publication Not required.