Abstract
Objective:
The factors associated with access-site failure following ultrasound-guided percutaneous access for aortic endograft procedures remain poorly characterized. We developed a prediction model to risk stratify patients for access site failure.
Methods
We performed a retrospective institutional review of consecutive patients who underwent endovascular aneurysm repair (EVAR), fenestrated EVAR (FEVAR), or thoracic EVAR (TEVAR) from 2014-2016. We excluded patients undergoing direct aortic access via sternotomy, and patients treated with physician modified endografts given reporting restrictions. Our primary outcome was groin access site failure, which included bleeding and thrombosis. An 8-point risk model was created for access site failure using multivariable fractional polynomials and internally validated using bootstrapping.
Results:
We identified 469 femoral arteries from 247 patients undergoing endovascular aortic repair procedures (EVAR 75%, FEVAR 8.0%, TEVAR 17%). Surgeons performed percutaneous access in 97.2% of the femoral arteries, with 99.6% ultrasound utilization. Twenty-seven (5.9%) access site failures occurred (17 bleeding, 10 thrombosis), all treated with groin cutdown, for a successful percutaneous femoral artery access rate of 94%. Of the 215 patients with attempted bilateral percutaneous access, 90% had successful bilateral access. However, FEVAR had lower rates of successful bilateral access (FEVAR 78 vs EVAR 91 vs TEVAR 94%, P = .03). Factors independently associated with percutaneous access site failure were femoral artery outer wall diameter (by mm: OR 0.003 [0.0002-0.1], P <.001), femoral artery stenosis >50% (OR 22.3 [2.7-183.2], P < .01), and urgent/emergent intervention (OR 3.6 [1.2-11.0], P = .03). A risk prediction model based on these criteria produced a C statistic of 0.89, a Hosmer-Lemeshow goodness of fit of 0.99, and a Brier score of 0.04. Excluding treatment for ruptured aneurysms, cutdown for access failure and planned initial groin cutdown resulted in longer post-op lengths of stay, and higher rates of access related readmission, return to operating room, groin infection, and myocardial infarction, compared to successful percutaneous access. There was no difference in major adverse events between planned initial groin cutdown and cutdown after failure, however the small number of patients in these two comparison groups limits the statistical power to detect a difference.
Conclusions:
Percutaneous ultrasound-guided access can be safely performed in almost all patients undergoing endovascular aortic procedures, but access site failures do occur. This risk score can help users select patients with high likelihood of success, identify patients who need close scrutiny with post-closure femoral duplex, and provide patient guidance about risk of unplanned groin cutdown.
Here is the edited TOC summary:
In this retrospective study of 247 patients who underwent percutaneous endovascular aortic repair, access site failures occurred in 5.9% of femoral arteries. Factors associated with access site failure included femoral artery outer wall diameter, femoral artery stenosis >50%, and urgent/emergent intervention.
Introduction
The use of percutaneous vascular closure devices to achieve hemostasis without the need for groin cutdown has allowed for the increasing use of percutaneous endovascular aortic aneurysm repair (EVAR).1–4 Traditionally, open femoral cutdowns were performed for insertion of large diameter sheaths. However, complications related to groin cutdown such as wound infection, lymphoceles, hematomas, excess pain, and delayed ambulation can occur, resulting in increased hospital length of stays and readmissions.5,6
Randomized and observational studies for percutaneous EVAR have shown both high success and superiority to femoral cutdown in terms of postoperative complications when used for standard infrarenal abdominal aortic aneurysm (AAA) devices.7,8 However there are limited data evaluating the outcomes of percutaneous EVAR for more complex aortic interventions and in patients with more challenging femoral anatomy. Furthermore, despite the hopes of circumventing access site complications with a completely percutaneous approach, bleeding and thrombosis still occur, necessitating femoral cutdown and arterial repair.
At our institution, we have adopted an ultrasound-guided percutaneous-first approach for almost all endovascular aortic cases including TEVAR and fenestrated EVAR for both aneurysm and dissection. We therefore conducted a retrospective institutional review of patients treated with percutaneous aortic endovascular procedures to determine factors independently associated with access site failure and to build a risk prediction model to help surgeons identify patients at high risk for access site failure.
Methods
Dataset and Patients
We retrospectively identified all consecutive patients (N=247) undergoing aortic endovascular procedures including endovascular aortic repair (EVAR), thoracic EVAR (TEVAR), and fenestrated EVAR (FEVAR) at our institution from January 1, 2014 to December 31, 2016. Patients treated with extension grafts for type Ia/Ib endoleaks or graft relining for type III endoleaks were also included. We excluded patients undergoing aortic endovascular repair with direct aortic access via sternotomy (N=2). Patients undergoing physician modified endografts were excluded given research reporting restrictions. This study was approved by the Institutional Review Board at Beth Israel Deaconess Medical Center and informed consent was waived given the retrospective nature of the design.
Baseline and Anatomic Characteristics
Patient demographics and anatomic characteristics were obtained via chart review. PL and MS performed all imaging review and anatomic measurements on preoperative computed tomography angiography (CTA) scans using the Conserus ™ Enterprise Viewing System (Change Healthcare, Nashville, TN). Degree of femoral stenosis was excluded from analysis for patients without contrast imaging (3.3%). Bilateral femoral artery measurements were performed if both groins were accessed for the procedure. We treated each artery individually, keeping baseline patient characteristics associated with each artery for analysis. In certain cases, percutaneous femoral artery access was performed via the proximal superficial femoral artery (SFA) (N=5) if patients had a prior common femoral artery (CFA) patch or bypass graft to or from the CFA. For these patients, SFA measurements were taken. Vessel wall diameter was determined by minimum outer wall diameter. We subjectively quantified the degree of calcification as none, minimal, moderate, and severe. Outer sheath diameters were determined based on data from the graft manufacturers. Femoral artery depth was measured from skin to anterior femoral wall.
Outcomes
The primary outcome was access site failure, defined as bleeding or thrombosis. Secondary outcomes included 30-day mortality, postoperative length of stay, discharge disposition, readmission, access related readmission, return to operating room (OR), groin infection, and myocardial infarction. For secondary outcomes, patients who were treated for ruptured aneurysms were excluded from analysis because the nature of ruptured aneurysms could overshadow the access-related outcomes (e.g. length of stay, myocardial infarction).
Technique
At our institution, a “percutaneous-first” approach is used for EVAR and femoral cutdown is reserved for femoral arteries that are nearly occluded on pre-operative imaging, or for patients with concomitant femoral aneurysms or symptoms of limb ischemia related to femoral artery stenosis. Ultrasound-guidance is used for all percutaneous access except for the rare circumstance of hemodynamic instability and an ultrasound is not readily available. Ultrasound is first performed in a cross-sectional view to characterize plaque and calcium. We then switch to a longitudinal view to visualize the entire length of the micropuncture needle enter the least calcified part of the artery in an area with minimal posterior plaque.9
Perclose Proglide suture-mediated closure devices (Abbott, Abbott Park, Illinois) are used in all cases. Two Perclose devices are placed in preclose fashion for all large sheaths (≥12Fr). We do not attempt to twist the Perclose device to 10 and 2 o’clock positions as we noted a higher failure rate anecdotally with this method and instead keep them both aligned at 12 o’clock. For smaller sheaths, one Perclose device is deployed in preclose or postclose fashion. Additional Perclose devices may be advanced over the wire and deployed if there is poor hemostasis. In the setting of access site failure, the sheath is replaced over the wire for hemostasis and a surgical cutdown is performed to repair the vessel.
We perform an on-table femoral duplex with color flow and pulse wave doppler for patients without palpable pedal pulses following deployment of Perclose devices to rule out vessel narrowing or thrombosis. After leaving the operating room, imaging is obtained for any concerns of arterial bleeding or thrombosis. Patients are seen for follow-up at 1 month where each femoral access site and distal pulse exams are assessed clinically. Additionally, a CTA that extends to the femoral vessels is obtained to evaluate the aortic repair and to identify access site complications.
Statistical Analysis
Continuous variables were presented as mean ± standard deviation. Categorical variables were presented as counts and percentages. Univariate differences between patients with and without access site failure were assessed using χ2 and Fisher’s exact tests for categorical variables and Student’s t-test or rank-sum for continuous variables where appropriate. All tests were two-tailed and p<0.05 was considered statistically significant.
Model Estimation
We used mixed effects logistic regression, allowing for random effects at the patient level to account for the fact that many patients had two femoral arteries assessed, to construct our risk score using the sequential Royston and Altman model-selection algorithm with multivariable fractional polynomials for the continuous variables using a threshold of <.05. This method tests various polynomial transformations for the continuous covariates in the model to determine the best fit, and drops nonsignificant covariates in a cyclical, backwards selection process. Our initial model included factors that were determined a priori on basis of their plausible associated with access site failure. This included gender, aspirin, anticoagulation, INR, BMI, narrowing >50%, femoral artery calcification, femoral calcification location, sheath diameter, vessel diameter, urgent/emergent case, and any prior groin interventions. The β coefficients were then used to assign point values to each of the variables included in the model.10
Model Performance and Validation
We assessed the discrimination, calibration and accuracy of our model using the C statistic of the receiver operating curve, the Hosmer-Lemeshow goodness-of-fit test, and the Brier score. The predictive ability of the risk model was then internally validated using a bootstrap sample with 200 replications.11,12 We applied our model to each of these data sets and calculated a C-statistic. The average of the differences between the C-statistic in the original data set and the bootstrapped data sets estimates the optimism of the C-statistic.13 The calibration slope was calculated by the average of the regression coefficients of the linear predictor from the original model applied to each bootstrapped data set. A slope of 1 indicated similar performance, whereas a slope <1 indicates worse discrimination and >1 indicates better discrimination. All statistical analysis was performed using Stata software (StataCorp, College Station, TX).
Results
A total of 469 femoral arteries were accessed in 247 patients undergoing endovascular aortic repair procedures; 300 (65.8%) for treatment of infrarenal AAA, 43 (9.4%) for juxtarenal AAA, 33 (7.2%) for thoracic aneurysms, 12 (2.6%) for dissection, 22 (4.8%) for penetrating aortic ulcers, and 13 (2.9%) for iliac aneurysms. The most common aortic device type was a standard EVAR (N=340, 75%), followed by TEVAR (N=79, 17%), FEVAR (N=36, 8.0%), and combined EVAR and TEVAR (N=1, 0.2%). The main body device was placed in the right groin in 331 of the 456 (72.6%) femoral arteries accessed. Twenty-eight (6.7%) of femoral arteries that underwent initial preclose technique had 1 additional Perclose device deployed at the end of the case, 3 (0.7%) had 2 additional devices, and 1 (0.2%) had 3 additional devices.
Of the 469 femoral arteries, 13 femoral arteries (2.9%) were deemed inappropriate for percutaneous access and treated with initial groin cutdown; 2 for concomitant femoral-femoral bypass, 1 for concomitant CFA aneurysm repair, 1 for a high CFA bifurcation, 2 for presence of prior CFA endarterectomy and patch angioplasty, and 7 for near occlusion of the CFAs. Percutaneous access was performed in the remaining 456 (97.2%) arteries. Nearly all arteries (N=454, 99.6%) were accessed under ultrasound guidance with no failures in sheath placement. The one patient with bilateral femoral arteries accessed without ultrasound presented with aortic rupture and the ultrasound was not readily available. Following percutaneous access, 27 (5.9%) had access site failures; 17 (3.7%) secondary to bleeding and 10 (2.2%) secondary to arterial thrombosis. This resulted in a successful percutaneous access rate of 94%. No failures occurred in the 5 patients who underwent SFA access.
Of the 247 total patients, 222 patients (90%) underwent an aortic endograft placement that required access to bilateral groins. Bilateral percutaneous access was attempted 215 of these patients (97%). Of the patients with attempted bilateral groin access, 194 patients (90%) had successful bilateral percutaneous access. Although there were no differences in access site failure among EVAR, TEVAR, or FEVAR for each individual artery studied, FEVAR resulted in lower rates of successful bilateral access rates (FEVAR 78 vs EVAR 91 vs TEVAR 94%, P = .03) (Table I).
Table I.
Successful Percutaneous Access Rates
| Successful Individual Groin Access N = 456 arteries (%) |
P value | Successful Bilateral Groin Access N = 215 patients (%) |
P value | |
|---|---|---|---|---|
| All Procedures | 94 | .36 | 90 | .03 |
| EVAR | 94 | 91 | ||
| TEVAR | 96* | 94# | ||
| FEVAR | 89 | 78 | ||
| Elective | 96 | .08 | 93 | <.01 |
| EVAR | 97 | 95 | ||
| TEVAR | 96* | 95# | ||
| FEVAR | 89 | 78 | ||
| Urgent/Emergent | 88 | .15 | 81 | .04 |
| EVAR | 86 | 78 | ||
| TEVAR | 96* | 93# | ||
| FEVAR | - | - |
EVAR, endovascular abdominal aortic aneurysm repair; TEVAR, thoracic endovascular aneurysm repair; FEVAR, fenestrated endovascular aneurysm repair.
Excludes diagnostic catheters (4 or 5 Fr).
Includes diagnostic catheters and single groin TEVAR
During the study period, access site failures only occurred after sheath removal and all resulted in surgical cutdown and arterial repair. No retroperitoneal hematomas occurred in the perioperative period. One late femoral pseudoaneurysm was found, occurred after discharge, and was not amenable to thrombin injection because of a wide neck and was therefore repaired in the operating room. No late access site stenosis or thrombosis was detected on 1-month CT imaging.
Of the 17 arteries complicated by uncontrolled bleeding, 16 (94%) were discovered intraoperatively. Preclose deployment failure occurred in 3 (17.6%) of the arteries due to high calcium burden. In all cases where bleeding was identified intra-operatively, wire access was maintained, and therefore sheath replacement for temporary hemostasis was possible prior to definitive surgical repair. In the 10 arteries complicated by thrombosis, 7 failures (70%) were discovered intraoperatively. Of the 3 failures discovered postoperatively, 2 did not undergo femoral duplex prior to leaving the operating room.
Patients with access site failure were more often female (52 vs 28%, P=.01), more likely to have documented peripheral vascular disease (30 vs 14%, P=.03) and had a higher preoperative INRs (1.3 vs 1.1 seconds, P=.03) (Table II). Prior ipsilateral femoral endarterectomy or presence of an ipsilateral bypass graft to or from the femoral artery was not associated with access site failure (Table III). However, patients with failure were more likely to undergo an urgent or emergent procedure (52 vs. 26%, P=<.01) for an uncontained rupture (37 vs 6%, P=<.001).
Table II.
Preoperative Characteristics of Patients Undergoing percEVAR with and without Access Site Failure
| No Failure N = 429 |
Failure N = 27 |
P value | |
|---|---|---|---|
| Age (years, mean ± SD) | 74.0 ± 10.3 | 73.9 ± 11.5 | .97 |
| Female | 28 | 52 | .01 |
| White | 80 | 78 | .79 |
| CAD | 36 | 37 | .94 |
| Prior MI | 25 | 30 | .59 |
| CABG | 11 | 15 | .49 |
| Hypertension | 83 | 81 | .87 |
| BMI > 40 | 6.1 | 15 | .08 |
| Diabetes | 19 | 22 | .69 |
| Hyperlipidemia | 76 | 56 | .02 |
| COPD | 17 | 26 | .27 |
| Stroke | 9.6 | 7.4 | .71 |
| PVD | 14 | 30 | .03 |
| History of smoking | 84 | 81 | .74 |
| Current Smoker | 30 | 44 | .11 |
| INR (seconds, mean ± SD) | 1.1 ± 0.4 | 1.3 ± 0.7 | .03 |
| Preoperative Medications | |||
| Aspirin | 70 | 61 | .29 |
| Plavix | 5 | 0 | .22 |
| Anticoagulation | 19 | 22 | .69 |
| Statin | 69 | 52 | .06 |
| Beta-blocker | 57 | 59 | .81 |
| ACEi | 33 | 33 | .96 |
ACEi, angiotensin-converting-enzyme inhibitor; BMI, body mass index; CABG, coronary artery bypass graft; COPD, chronic obstructive pulmonary disease; INR, international normalized ratio; MI, myocardial infarction; PVD, peripheral vascular disease; SD, standard deviation. Values are in % unless otherwise indicated. Bolded values indicate statistical significance at P<.05.
Table III.
Prior Surgical History and Surgical Status of Patients Undergoing percEVAR with and without Access Site Failure
| No Failure N = 429 |
Failure N = 27 |
P value | |
|---|---|---|---|
| Prior Femoral Artery Surgery | .26 | ||
| Endarterectomy | 2.3 | 7.4 | |
| Bypass | 0.7 | 0.0 | |
| Prior Percutaneous Groin Access | 12 | 11 | .93 |
| Prior AAA repair | 0.7 | ||
| None | 93 | 96 | |
| EVAR | 4.7 | 3.7 | |
| Open | 2.8 | 0.0 | |
| Procedure Urgency | <.01 | ||
| Elective | 74 | 48 | |
| Urgent/Emergent | 26 | 52 | |
| Degree of Rupture | <.001 | ||
| Not ruptured | 85 | 59 | |
| Contained | 9 | 4 | |
| Uncontained | 6 | 37 |
AAA, abdominal aortic aneurysm; EVAR, endovascular aortic aneurysm repair. Values are in % unless otherwise indicated.
Anatomic Factors
Average sheath and femoral artery diameters are listed in Table IV. Patients with access site failure were more likely to have a smaller femoral artery outer diameter (7.9 vs 10.3mm, P=<.001). There was a left skew in femoral vessel outer diameter with a predominance of access site failures occurring in smaller diameter vessels (Figure 1). Although most patients had femoral arteries with less than 25% stenosis, higher degrees of femoral artery stenosis were associated with higher rates of access site failure (Figure 2). Femoral artery depth was not associated with failure.
Table IV.
Procedural Characteristics
| No Failure N = 429 |
Failure N = 27 |
P value | |
|---|---|---|---|
| Sheath (Fr) | 15.3 ± 4.7 | 15.6 ± 3.6 | 0.75 |
| Sheath outer diameter (mm) | 5.7 ± 1.6 | 5.9 ± 1.4 | 0.56 |
| Access vessel diameter (mm) | 10.3 ± 2.3 | 7.9 ± 1.8 | <.001 |
| Sheath/Artery Ratio | 2.0 ± 1.0 | 1.4 ± 0.6 | <.01 |
| Femoral calcification | <.001 | ||
| None | 27 | 24 | |
| Minimal | 44 | 12 | |
| Moderate | 26 | 40 | |
| Severe | 2.9 | 24 | |
| Calcification location | 0.85 | ||
| None | 28 | 24 | |
| Anterior | 1.4 | 4.0 | |
| Posterior | 47 | 52 | |
| Lateral | 9.8 | 8.0 | |
| Concentric | 14 | 12 | |
| Femoral Artery Narrowing | <.001 | ||
| 0-10% | 48 | 18 | |
| 10-25% | 43 | 36 | |
| 25-50% | 8.1 | 18 | |
| 50-75% | 1.0 | 18 | |
| >75% | 0.3 | 9.1 | |
| Femoral Artery Depth (mm) | 29.5 ± 11.2 | 28.8 ± 16.1 | 0.75 |
| Moderate/Severe Iliac Tortuosity | 41 | 24 | 0.09 |
| ACT (seconds) | 240.4 + 26.3 | 246.0 + 32.1 | 0.36 |
| Sheath in artery time (minutes) | 103.8 + 54.7 | 102.3 + 64.8 | 0.90 |
Fr, French; mm, millimeter. Values are in % unless otherwise indicated.
Figure I.
Femoral artery outer wall diameter in millimeters (horizontal axis) with the absolute number of treated arteries within each access vessel diameter range (left axis) and the proportion of arteries with percutaneous failure within each group (right axis and above bars).
Figure II.
Percent femoral artery stenosis (horizontal axis) with the absolute number of treated arteries within each femoral artery stenosis range (left axis) and the proportion of arteries with percutaneous failure within each group (right axis and above/within bars).
Smaller sheath to artery ratio (1.4 vs 2.0, P<.01) was also associated with access site failure. However, sheath size alone was not associated with access site failure (5.9 vs 5.7mm, P=.56). The majority of femoral arteries had minimal (42.3%) or moderate (26.6%) femoral artery calcification. Calcification location was most commonly found at the posterior aspect of the artery (47.3% overall). Although more severe femoral artery calcification was associated with higher rates of access failure, there was no significant association with the location of femoral artery calcification and access site failure.
Secondary Outcomes
Arteries accessed in the setting of contained (n=41) or free aortic rupture (n=35) were excluded from secondary outcome analysis. Patients who underwent cutdown for access failure or had planned initial cutdown had significantly more post-operative major adverse events compared to patients with successful percutaneous groin access. Compared to patients with successful percutaneous access, groin cutdown for any reason resulted in longer post-op lengths of stay and higher rates of access related readmission, discharge to rehab, return OR , groin infection, and myocardial infarction (Table V). There was no difference in terms of 30-day mortality or any readmission. When comparing patients with planned initial cutdown and cutdown after failure, there was no difference in major adverse events. However, given the small number of patients in both groin cutdown cohorts, the risk for type II error is high and the power of this study to detect a significant difference is low.
Table V.
Post-op Complications and Disposition
| Planned Initial Cutdown N = 12 |
No Cutdown or Access Failure N = 365 |
Cutdown after Failure N = 16 |
P value | |
|---|---|---|---|---|
| 30 day Mortality | 0.0 | 1.6 | 0.0 | .79 |
| Postop LOS (days) | 6.4 ± 2.5 | 2.7 ± 5.4 | 4.6 ± 3.4 | .02 |
| Disposition | <.001 | |||
| Home | 17 | 75 | 7.3 | |
| Home w/VNA | 25 | 17 | 42 | |
| Rehab | 63 | 7.3 | 25 | |
| Readmission | 17 | 20 | 33 | .44 |
| Access related readmission | 8.3 | 0.9 | 20 | <.001 |
| Return to OR | 8.3 | 0.6 | 19 | <.001 |
| Groin Infection | 8.3 | 0.3 | 6.3 | <.001 |
| Myocardial infarction | 0.0 | 0.3 | 6.3 | <.01 |
LOS, length of stay; OR, operating room; VNA, visiting nursing association. Values are in % unless otherwise indicated.
TEVAR and FEVAR procedures require larger sheaths and can add an additional level of technical complexity to endovascular aortic repair. After adjustment for complex endovascular aortic procedures, any groin cutdown (either planned or unplanned) compared to successful percutaneous access still resulted in higher odds of access related readmission (odds ratio (OR) 20.3 [95% CI 2.8-373.5], P < .001), failed disposition home (OR 11.5 [4.5-29.3], P < .001), and groin infection (OR 32.2 [2.8-373.5], P < .01]. However, groin cutdown was no longer independently associated with post-operative myocardial infarction (OR 15.2 [0.9-253.5], P = .06) after adjustment.
Risk Model
Factors independently associated with access site failure included femoral vessel outer diameter (per millimeter increase: OR 0.003 [.0002-.1]; P<.001), vessel stenosis greater than 50% (OR 22.3 [2.7-183.2], P <.01), and urgent/emergent case status (OR 3.6 [1.2-11.0], P=.03) (Table VI). These factors were used to construct our risk model. The model resulted in a favorable C statistic of 0.89, goodness of fit of 1.00, and Brier score of .04. β coefficients were used to assign point values to each of the three variables to create a risk model with points values ranging from 0 to 8 (Table VII). The model performed well in the internal validation sample, with an optimism-corrected C statistic of 0.88 and a calibration slope of 1.04.
Table VI.
Model Selection Characteristics
| Characteristic | Odds Ratio [95% Confidence Interval] |
P-Value |
|---|---|---|
| Vessel Diameter | .003 [.0002-.1] | <.001 |
| Vessel Stenosis >50% | 22.3 [2.7-183.2] | <.01 |
| Urgent/Emergent | 3.6 [1.2-11.0] | .03 |
Table VII.
Risk Score Criteria
| Points | |
|---|---|
| Procedure Urgency | |
| Elective | 0 |
| Urgent/Emergent | 1 |
| Femoral Stenosis | |
| <50% | 0 |
| >50% | 3 |
| Femoral Vessel Outer Diameter (mm) | |
| >12 | 0 |
| 10-12 | 2 |
| 8-10 | 3 |
| <8 | 4 |
Discussion
Percutaneous ultrasound-guided access can be safely performed in almost all patients undergoing endovascular aortic procedures. We found that smaller femoral vessel diameters, femoral artery stenosis greater than 50%, and urgent/emergent cases were independently associated with percutaneous access site failure for endovascular aortic procedures. Given these findings, we developed a model to identify a cohort of patients who are at high-risk for access site failure and found that our model accurately stratifies patients on just three preoperative factors, which can be easily be implemented prior to a case. Successful percutaneous groin access is associated with lower rates of major adverse events compared to patients undergoing planned or unplanned groin cutdown.
Use of percutaenous EVAR has rapidly grown over the past few years, with an increasing number of physicians now adopting this technique as their preferred approach, leading to upwards of 64% of all EVARs performed percutaneously nationwide.14 As widespread experience and popularity with percutaneous EVAR continues to grow, the limits of percutaneous access in more anatomically challenging femoral vessels will be tested. Despite a high technical success rate with ultrasound-guided percutaneous access at our institution and development of smaller and more flexible devices, access site failure after percutaneous access continue to occur, resulting in unplanned groin cutdown and added risk to the patient. This highlights the ongoing need to perform these procedures in an operating room with the appropriate instruments, lighting, and sterile environment.
Four years after the introduction of the ultrasound-guided percutaneous EVAR at our institution, we rapidly reached a 92.3% attempted percutaneous access rate for all elective, standard EVAR cases.5 In this study, we report a 97.2% attempted percutaneous access rate, while expanding the technique to larger devices (FEVAR and TEVAR), ruptured aneurysms, and to patients with more heavily diseased femoral vessels. Our percutaneous-first approach and willingness to attempt percutaneous access in potentially hostile groins has resulted in slightly higher rates of access site failures compared to our prior study (6% vs 4%). However, although we push the limits in terms of attempting percutaneous access, our results compare favorably to similar reports in the literature, with higher failure rates ranging from 6.5 to 12% for standard EVAR and up to 10% in more complex TEVAR, FEVAR, and branched stent graft devices.8,15,16 Our low access failure rates may be due to the fact that although we attempt percutaneous access in quite challenging situations, this is balanced by the resultant increase in percutaneous experience and an overall high rate of successful percutaneous access. We also report no access failures in patients in whom we performed SFA access, however we only recommend percutaneous SFA access in carefully selected patients who have non-diseased SFAs with adequate diameters.
In general, success with percutaneous access is dependent on the user’s experience with routine ultrasound guidance for percutaneous arterial access.17 The benefit of ultrasound-guided access is not only seen in percutaneous aortic cases but has also been proven in lower extremity cases.9,18,19 In this study we report a 99.6% ultrasound usage rate with 100% success in initial arterial entry and all access site failures occurring after sheath removal. In our study, access site failure following percutaneous EVAR was independently associated with the vessel diameter, degree of femoral stenosis, and having an urgent or emergent case. Although limited to standard infrarenal devices, other studies have found similar risk factors associated with access site failure, including small femoral access diameter, female gender, age, and severe arterial calcifications.3,16,20 We also found that higher degrees of femoral calcification was associated with access site failure on univariate analysis. However, in our risk model, calcification was removed because severity of calcification was found to be collinear with vessel stenosis and percent arterial stenosis was a stronger predicter of access failure. A risk model containing femoral calcification (shown in supplementary tables 1 and 2) may be useful in patients with only non-contrast imaging of the femoral vessels for which the degree of femoral stenosis cannot be accurately characterized.
The implications of access failure following percutaneous EVAR include both patient safety and financial concerns. Patients who undergo groin cutdown following access failure have longer lengths of stay, are more likely to be discharged to rehab, and have higher rates of return to the operating room and access related readmission. Although we found that access site complications are often identified intraoperatively, some require return to the operating room, reintubation, and reintervention. This highlights the importance of close post-operative monitoring in patients undergoing percutaneous EVAR to avoid the added risks of a reoperation. Additionally, given the cost of closure devices, studies have shown that the financial benefit of percutaneous EVAR over cutdown EVAR is only maximized when percutaneous EVAR complication rates are low.21
We found that our risk model performed and correlated well with access site failure. Therefore, we believe this risk score will be helpful for surgeons in the early part of the percutaneous EVAR learning curve to avoid performing percutaneous EVAR in patients at high risk for access site failure. For more experienced percutaneous EVAR users, this risk score will help identify patients who will need more careful access site assessment following closure device deployment to avoid reoperations, even in the setting of emergent ruptured AAAs. As most of our access failures were recognized intraoperatively, we believe that the rate of our reoperations and return to the operating room can be lowered by having a low threshold for using intraoperative femoral duplex in higher risk patients. We now perform routine intraoperative duplex in patients with absent pedal pulses or calculated risk score ≥4. We have also implemented this risk score to assist in preoperative discussions with patients regarding the likelihood of having a groin incision.
The results of this study must be interpreted in the context of its design and the database used. Based on the non-randomized retrospective nature of this study, the treating surgeon determined the selection for percutaneous access over groin cutdown. Given our preference for percutaneous EVAR, this study lacked a large comparison group with initial open groin exposure for access. The experience and high usage of ultrasound-guided percutaneous access may also not reflect a typical vascular practice. However, we feel that this risk score can still be beneficial for surgeons who are initially learning percutaneous EVAR, to identify select cases with a high probability for success.
Conclusion
Percutaneous ultrasound-guided access can be safely performed in almost all patients undergoing endovascular aortic procedures. However, access site failures can occur, resulting in unplanned groin cutdowns. Application of this risk model can help accurately identify patients who will are at highest risk for access site failure.
Supplementary Material
Table VIII.
Access Failure Risk, Frequency, and Cumulative Frequency by Risk Score
| Total Points | Predicted Risk of Access Failure (%) |
Frequency (%) | Cumulative Frequency (%) |
|---|---|---|---|
| 0 | 0.0 | 13.4 | 13.4 |
| 1 | 0.0 | 5.0 | 18.3 |
| 2 | 0.4 | 24.4 | 43.0 |
| 3 | 1.2 | 32.6 | 75.5 |
| 4 | 3.8 | 16.5 | 91.9 |
| 5 | 10.8 | 6.1 | 98.0 |
| 6 | 27.3 | 0.5 | 98.4 |
| 7 | 53.9 | 1.6 | 100 |
| 8 | 78.3 | 0.0 | 0.0 |
JVS-D-18-01294R2, Preoperative Risk Score for Access Site Failure in Ultrasound-Guided Percutaneous Aortic Procedures
Type of Research: Retrospective cohort study
Key Findings: In 247 patients who underwent percutaneous endovascular aortic repair, access site failures occurred in 5.9% of femoral arteries. Factors associated with access site failure included femoral artery outer wall diameter, femoral artery stenosis >50%, and urgent/emergent intervention.
Take Home Message: Percutaneous access site complications for endovascular aortic repairs are associated with femoral artery outer wall diameter, femoral artery stenosis >50%, and urgent/emergent intervention. A risk score to predict access complications is proposed.
Acknowledgments
PL and TO supported by the Harvard-Longwood Research Training in Vascular Surgery NIH T32 Grant 5T32HL007734-22
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Watelet J, Gallot JC, Thomas P, Douvrin F, Plissonnier D. Percutaneous Repair of Aortic Aneurysms: A Prospective Study of Suture-Mediated Closure Devices. Eur J Vasc Endovasc Surg. 2006; [DOI] [PubMed] [Google Scholar]
- 2.Lee WA, Brown MP, Nelson PR, Huber TS. Total percutaneous access for endovascular aortic aneurysm repair (“Preclose” technique). J Vasc Surg. 2007; [DOI] [PubMed] [Google Scholar]
- 3.Manunga JM, Gloviczki P, Oderich GS, Kalra M, Duncan AA, Fleming MD, et al. Femoral artery calcification as a determinant of success for percutaneous access for endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2013; [DOI] [PubMed] [Google Scholar]
- 4.Van Orden K, Farber A, Schermerhorn ML, Goodney PP, Kalish JA, Jones DW, et al. Local anesthesia for percutaneous endovascular abdominal aortic aneurysm repair is associated with fewer pulmonary complications. J Vasc Surg. 2018; [DOI] [PubMed] [Google Scholar]
- 5.Bensley RP, Hurks R, Huang Z, Pomposelli F, Hamdan A, Wyers M, et al. Ultrasound-guided percutaneous endovascular aneurysm repair success is predicted by access vessel diameter. J Vasc Surg. 2012;55(6):1554–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Buck DB, Karthaus EG, Soden PA, Ultee KHJ, Van Herwaarden JA, Moll FL, et al. Percutaneous versus femoral cutdown access for endovascular aneurysm repair. J Vasc Surg. 2015; [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Nelson PR, Kracjer Z, Kansal N, Rao V, Bianchi C, Hashemi H, et al. A multicenter, randomized, controlled trial of totally percutaneous access versus open femoral exposure for endovascular aortic aneurysm repair (the PEVAR trial). J Vasc Surg. 2014; [DOI] [PubMed] [Google Scholar]
- 8.De Souza LR, Oderich GS, Banga PV, Hofer JM, Wigham JR, Cha S, et al. Outcomes of total percutaneous endovascular aortic repair for thoracic, fenestrated, and branched endografts. J Vasc Surg. 2015; [DOI] [PubMed] [Google Scholar]
- 9.Lo RC, Fokkema MTM, Curran T, Darling J, Hamdan AD, Wyers M, et al. Routine use of ultrasound-guided access reduces access site-related complications after lower extremity percutaneous revascularization. J Vasc Surg. 2015; [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sullivan LM, Massaro JM, D’Agostino RB. Presentation of multivariate data for clinical use: The Framingham Study risk score functions. Stat Med 2004; [DOI] [PubMed] [Google Scholar]
- 11.Steyerberg EW, Eijkemans MJC, Harrell FE Jr, Dik J, Habbema F. Sensible Strategy for Prognostic Modeling with Logistic Regression? Prognostic Modeling with Logistic Regression Analysis: In Search of a Sensible Strategy in Small Data Sets. Habbema Med Decis Mak. 2001; [DOI] [PubMed] [Google Scholar]
- 12.Steyerberg EW, Harrell FE, Borsboom GJ, Eijkemans MJ, Vergouwe Y, Habbema JD. Internal validation of predictive models: efficiency of some procedures for logistic regression analysis. J Clin Epidemiol. 2001; [DOI] [PubMed] [Google Scholar]
- 13.Steyerberg EW, Vickers AJ, Cook NR, Gerds T, Obuchowski N, Pencina MJ, et al. Assessing the performance of prediction models : A framework for some traditional and novel measures. Epidemiology. 2010; [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Siracuse JJ, Farber A, Kalish JA, Jones DW, Rybin D, Doros G, et al. Comparison of access type on perioperative outcomes after endovascular aortic aneurysm repair. J Vasc Surg. 2018; [DOI] [PubMed] [Google Scholar]
- 15.Rijkée MP, Statius Van Eps RG, Wever JJ, Van Overhagen H, Van Dijk LC, Knippenberg B. Predictors of failure of closure in percutaneous EVAR using the prostar XL percutaneous vascular surgery device. Eur J Vasc Endovasc Surg. 2015; [DOI] [PubMed] [Google Scholar]
- 16.Mousa AY, Campbell JE, Broce M, Abu-Halimah S, Stone PA, Hass SM, et al. Predictors of percutaneous access failure requiring open femoral surgical conversion during endovascular aortic aneurysm repair. J Vasc Surg. 2013; [DOI] [PubMed] [Google Scholar]
- 17.Bechara CF, Barshes NR, Pisimisis G, Chen H, Pak T, Lin PH, et al. Predicting the learning curve and failures of total percutaneous endovascular aortic aneurysm repair. J Vasc Surg. 2013; [DOI] [PubMed] [Google Scholar]
- 18.Starnes BW, O’Donnell SD, Gillespie DL, Goff JM, Rosa P, Parker MV., et al. Percutaneous arterial closure in peripheral vascular disease: A prospective randomized evaluation of the Perclose device. J Vasc Surg. 2003; [DOI] [PubMed] [Google Scholar]
- 19.Kalish J, Eslami M, Gillespie D, Schermerhorn M, Rybin D, Doros G, et al. Routine use of ultrasound guidance in femoral arterial access for peripheral vascular intervention decreases groin hematoma rates. J Vasc Surg. 2015; [DOI] [PubMed] [Google Scholar]
- 20.Eisenack M, Umscheid T, Tessarek J, Torsello GF, Torsello GB. Percutaneous endovascular aortic aneurysm repair: a prospective evaluation of safety, efficiency, and risk factors. J Endovasc Ther. 2009; [DOI] [PubMed] [Google Scholar]
- 21.Uhlmann ME, Walter C, Taher F, Plimon M, Falkensammer J, Assadian A. Successful percutaneous access for endovascular aneurysm repair is significantly cheaper than femoral cutdown in a prospective randomized trial. Journal of Vascular Surgery. 2018. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.