Increased Vascular Access Complications in Patients with Renal Dysfunction Undergoing Percutaneous Coronary Procedures Using Arteriotomy Closure Devices

ABSTRACT: Background. Arteriotomy closure device (ACD) use has increased following percutaneous transfemoral coronary procedures (PTCP). However, their safety in patients with chronic kidney disease (CKD) is not known. Therefore, we evaluated the complication rates of ACD among patients with CKD. Methods. Six-hundred ten consecutive patients who underwent PTCP and ACD were retrospectively studied. Patients were grouped according to their creatinine clearance (CrCl in ml/min/1.73 m2) calculated by the Cockcroft-Gault formula using the National Kidney Foundation classification system; Stage I (CrCl ≥ 90); Stage II (60–89); Stage III (30–59); Stage IV (15–29); and Stage V (≤ 15). The primary endpoint was the combined incidence of pseudo-aneurysm, retroperitoneal hematoma, femoral artery thrombosis, surgical vascular repair, and groin infection. Results. Among 610 patients 283 (46%) underwent PCI. The primary endpoint was seen in 66 (10.8%) patients. Univariate predictors of primary outcome were lower CrCl (p < 0.001), and presence of peripheral vascular disease (p = 0.03). There was an inverse relationship between CrCl and complication rate. CKD was the strongest independent multivariate predictor for the primary endpoint (OR 1.032; 95% CI 1.019–1.046; p < 0.0001), driven by higher infection (p < 0.0001), thrombosis (p = 0.003) and hematoma (p = 0.007). Conclusions. Renal function appears to be significantly associated with vascular access-site complications. Worsening renal function is associated with higher vascular access site complications, largely driven by an increased infection rate.

J INVASIVE CARDIOL 2010;22:8–13

Key words: kidney disease, cardiac catheterization, closure devices, vascular complications

Cardiovascular disease is a major cause of morbidity and is responsible for approximately 45% of all-cause mortality in patients with chronic kidney disease (CKD).1–3 CKD, as measured by the glomerular filtration rate (GFR), has been shown to be an independent predictor of adverse cardiovascular events including acute myocardial infarction and cardiac arrest.4,5
Diagnostic and therapeutic modalities are of increasing importance in patients with CKD who suffer from cardiovascular disease. Therefore, these patients represent a growing number of patients undergoing coronary revascularization. It is well established that patients with CKD undergoing coronary revascularization, whether coronary artery bypass graft (CABG) surgery or percutaneous transfemoral coronary procedures (PTCP), are at increased risk for post-procedure complications and increased mortality.6–8 These adverse clinical outcomes progressively increase as renal function worsens.6–8

The usage of arteriotomy closure devices (ACD) after PTCP as an alternative to traditional mechanical compression is on the rise.9 These devices have been developed with the intent to reduce hemostasis time, shorten bed rest and decrease hospital length of stay.10–14 The safety of these devices remains controversial, with reports of overall major complication rates at approximately 0.5–4%. However, most ACD trials have excluded patients with CrCl ≤ 60 ml/min, thus the safety data on the use of ACD in patients with CKD are very limited. The objective of this study was to determine whether ACD could be safely used in a high-risk population, specifically patients with CKD undergoing any percutaneous transfemoral diagnostic and/or interventional procedures.

Methods

We retrospectively studied data from 610 patients who underwent PTCP and ACD placement at our institution between January 1, 2005 and December 31, 2005. Diagnostic coronary catheterization and PTCP were performed according to standard guidelines. Unless contraindicated, all patients undergoing PTCP received aspirin, clopidogrel, weight-adjusted heparin and eptifibatide as needed.15

Weight-adjusted unfractionated heparin dosing was used to achieve an activated clotting time (ACT) using the Hemochron device (International Technidyne, Piscataway, New Jersey) of approximately 200–250 seconds for patients receiving glycoprotein (GP) IIb/IIIa inhibitors and > 300 seconds for all other patients.16,17 Patients in this study received GP IIb/IIIa receptor inhibitors according to usual protocol with eptifibatide including renal doses. Post-PCI patients were treated with aspirin (81 mg to 325 mg/day) and clopidogrel (300 or 600 mg as a loading dose followed by 75 mg/day) if stents were placed.

The cardiologist performing the procedure chose the method of arterial access management and type of ACD used. The ACD placement was performed by physicians trained in their use. The ACD was placed only after a femoral angiogram was performed via the arterial sheath. Patients with inappropriate vessel size (< 5 mm in diameter), high femoral arteriotomy site, excess calcification or plaque formation, presence of severe peripheral vascular disease, and arteriotomy sites in the SFA or profunda artery were excluded from ACD usage. Diabetic patients routinely received one dose of prophylactic antibiotic, either cephazolin 1 gm or vancomycin 1 gm IV prior to closure. All sheaths were removed soon after the procedure. Ambulation was initiated 2 hours for diagnostic and 4 hours for interventional cases after the ACD was placed. Access-site evaluation was routinely done after the procedure and before discharge.

Closure devices used. The closure devices used were the Angio-Seal® Vascular Closure Device (St. Jude Medical, Minnetonka, Minnesota) or The Perclose® suture device (Perclose, Inc., Redwood City, California). The Angio-Seal Vascular Closure Device VIP Platform consists of the Angio-Seal VIP device, an insertion sheath, an arteriotomy locator (modified dilator) and a guidewire. The Angio-Seal VIP device is composed of an absorbable collagen sponge and a specially designed absorbable polymer anchor that are connected by an absorbable self-tightening suture. The device seals and sandwiches the arteriotomy between its two primary members, the anchor and collagen sponge. Hemostasis is achieved primarily by the mechanical means of the anchor-arteriotomy-collagen sandwich, which is supplemented by the coagulation inducing properties of the collagen. The Angio-Seal Vascular Closure Device components are not made from latex rubber.18 The Perclose suture device incorporates two components: a sheath holding one or two pairs of needles connected by a suture loop and a rotating barrel used to facilitate the positioning of the device before needle deployment and to guide the needles during their travel through the subcutaneous track.18 The device uses a percutaneous means to suture and close the arterial site. Both devices are approved by the U.S. Food and Drug Administration.18

Estimation of Glomerular Filtration Rate (eGFR)

The patient’s creatinine Clearance (CrCl) in mL/min was calculated using the Cockcroft-Gault formula for eGFR estimation: (140 - age) x (weight) x (0.85 if female) / (72 x SCr). This formula has been appropriately validated and accurately predicts basal eGFR utilizing serum creatinine and the patient’s body weight.19–21 The kidney disease outcome quality initiative in 2001 developed guidelines that defined CKD as structural or functional abnormalities of the kidney, demonstrated most often by persistent albuminuria, with or without decreased eGFR (< 60 ml/min) or decreased eGFR, with or without other evidence of kidney damage. CKD was classified as per the National Kidney Foundation Classification as Stage I (CrCl ≥ 90), Stage II (60–89), Stage III (30–59), Stage IV (15–29) and Stage V (≤ 15, end-stage renal disease).21

Study endpoints. ACD-related complications were defined as unexpected adverse events that occurred during hospital stay. Data on complications for these patients were retrieved by retrospective review of the medical charts by investigators who were blinded to the renal function of the patient. The primary endpoint was cumulative in-hospital incidence of major vascular complication defined as bleeding complications, femoral artery thrombosis, groin infection and dissection requiring surgical vascular repair. Institutional review board approval was obtained for this study.

The complications of ACD placement were defined as any of the following:

1. Bleeding complications. Swelling > 6 cm or access site bleeding and/or (> 3g Hgb drop or > 10% drop in hematocrit) requiring transfusion or retroperitoneal bleeding confirmed on CT scan, including those that prolonged hospitalization and that required surgical intervention. Bleeding complications were defined as per the TIMI criteria.

2. Infection. Erythema, swelling and purulent drainage from the access site with or without systemic manifestations of fever and bacteremia. An ID specialist was consulted on all the patients who had infection and IV antibiotics were given.

3. Dissection or femoral artery thrombosis. Loss of peripheral pulses and thrombosis or dissection diagnosed by duplex ultrasound and requiring intervention.

4. Device failure. Any circumstance in which the device did not achieve hemostasis.

Statistical analysis. Statistical analysis was performed using a standard statistical software package (SPSS for Windows, version 16; SPSS, Inc., Chicago, Illinois). Continuous variables were expressed as the mean ± standard deviation. Normally distributed variables were compared by ANOVA. Categorical variables were expressed as a percentage of the total sample and compared using the chi-square test or Fischer’s exact test when appropriate. Risk factors for complications were expressed by univariate odds ratios. Multivariate logistic regression analysis was performed to determine the independent predictors for the development of complications. A multivariate logistic regression model including age, gender, diabetes, CKD stage, closure type and smoking was used. A p-value of < 0.05 was considered statistically significant.

Results

Tables 1 and 2 describe the baseline clinical and procedural characteristics of the study population classified according to the eGFR. Among 610 patients studied, 105 (17.2%) were in Stage I, 260 (42.6%) Stage II, 192 (31.4%) Stage III, 36 (5.9%) Stage IV, and 17 (2.8%) in Stage V kidney function. Two hundred eighty-three patients (46%) underwent interventional procedures, and 522 (85%) had a collagen device placed. Two hundred forty-five patients (67%) had CrCl < 60 ml/min; these patients tended to be older (mean age = 71 ± 10 years; p = 0.01), and were more likely to have a higher prevalence of hypertension and diabetes. There was a slightly higher interventional procedure rate in the group with CrCl > 60 ml/min (55% vs. 44%; p = 0.07). There were no significant differences among the groups in regard to smoking, body mass index (BMI) and presence of peripheral vascular disease (PVD). There was a trend toward a decrease in platelet counts as the renal function declined, with Stage ≥ IV patients having platelets < 200,000 (p < 0.001). No significant differences were noted among the groups based on the type of closure device deployed or the usage of GP IIb/IIIa inhibitors.

Vascular complications were seen in 66 (10.8%) patients. As shown in Figures 1 and 2, bleeding complications were seen in 37 (6%), and groin infections in 11 (1.7%) of the patients in the study. Ten (1.6 %) patients developed femoral artery dissection, 5 (0.8%) developed femoral artery thrombosis, and device malfunction was seen in 3 (0.5%) patients. There was a significant inverse relation between renal function and incidence of all complications, with rates of 53% seen in patients with Stage V kidney function (p < 0.001).

In an effort to further assess the relation between CKD, procedure type and complication rate post PTCP, we classified the entire cohort into four groups. Groups A and B were patients with CrCl > 60 mL/min who underwent diagnostic or interventional procedures, and Groups C and D were patients with CrCl < 60 mL/min who underwent a diagnostic or interventional procedure, respectively. In this classification, patients with CrCl < 60 mL/min comprised 242 (40%) of the total cohort. As seen with our primary analysis, despite undergoing fewer interventional procedures, Group D had the highest overall complication rate of 20% when compared to Groups A, B and C (4, 8, 14 respectively; p < 0.001; Figure 4). This high complication rate was driven mainly by bleeding complications (14% in Group D vs. 1.2%, 5%, 6% in groups A, B and C; p < 0.001) and higher infection rates (3.5% vs. 0.1%, 0.5, 5% in Groups A, B and C; p = 0.004) (Figures 5 and 6).

Univariate predictors of primary outcome are listed in Table 3, with a presence of CKD as the strongest predictor (odds ratio [OR] 2.05; p < 0.0001). Multivariate logistic regression models for determining predictors of complications are shown in Table 4. The presence of CKD (OR 2.11; p < 0.0001) remained the strongest predictor of the endpoint driven mainly by higher infection rates, as seen in Figure 3 (χ2 = 31.9; p < 0.0001), followed by femoral artery thrombosis (χ2 = 8.9; p = 0.003) and bleeding complication (χ2 = 3.18; p = 0.004).

Discussion

The aim of this study was to define the association between severity of CKD and the rate of vascular access complications in patients undergoing PTCP using ACD. The results of our study show an overall complication rate of 10.8%, which is higher than that in previous studies. This discrepancy is most likely due to high percentage of CKD patients, with 40% of the study population having CKD ≥ III. In a recently published study from the Mayo Clinic of 17,901 patients who underwent PCI, complication rates ranged from 3.5–8.4%. However, patients with moderate-to-severe renal impairment were underrepresented, comprising only 3.5% of the entire study population.24 In our study, there was a significant relationship between renal dysfunction and the incidence of vascular complications following ACD, mainly due to higher bleeding and infection rates in those with severe renal dysfunction reaching up to 20% in patients with CrCl < 60 mL/min who underwent interventional procedures. These findings may have important clinical implications, given the increase in the prevalence of CKD over time in the United States.25,26

Potential Explanation of Our Findings

As in other studies, bleeding was the most frequent complication noted in our study population, with an incidence rate of 6%. Further analysis revealed an increasing incidence of bleeding complications as the stage of CKD increased, with a complication rate of 30% in patients with Stage ≥ III. This observation is likely a result of uremia-induced platelet dysfunction, which is commonly seen in patients with advanced renal impairment. The platelet dysfunction characteristic of uremia is multifaceted. Platelet count is usually within the normal range or slightly low in patients with uremia. It has been suggested that these patients have a complex platelet dysfunction and an abnormal platelet-endothelial vessel wall interaction.26 Radioligand studies have indicated that the binding of fibrinogen to ADP-stimulated platelets in uremic media is impaired. Notably, the ability of the vessel wall to generate the potent antiaggregatory substance prostacyclin (prostaglandin I2) increases in uremia; moreover, endothelial cells seem to generate an abnormal complex of coagulation factor VIII (antihemophilic factor) and Von Willebrand factor (vWF). Finally, the largest polymers of vWF, which are primarily responsible for the adhesion process, are deficient in patients with uremia, although the serum level of vWF in these patients is usually high or within the normal range.26,27

Infectious complications are an uncommon but serious complication of ACD use, with a reported incidence of 0– 5.1%.28,31 Infectious complications were seen in 11 patients (1.6%) in our study and similar to the trend seen with bleeding complication, incidence rates increased significantly as the renal function declined. Bleeding complication at the access site and the presence of foreign material in the intravascular space and arterial wall likely serve as a nidus for subsequent infection. Uremia causes immune dysfunction and increased susceptibility to infection through a variety of mechanisms.32,33 Patients on hemodialysis often display compromised neutrophil and T-cell function as well as diminished antibody production. The diminished neutrophil function may in part be due to the use of bioincompatible dialysis membranes, which result in impaired adherence and attenuated responses to phagocytic stimuli. In addition to these mechanisms, some of the immune deficits may also be explained by the presence of elevated endogenous glucocorticoids present in patients with renal impairment.32,33 Neutrophil impairment is also associated with diabetes, a major cause of CKD, and correlates directly to the level of hyperglycemia. Moreover, diabetics often have poor peripheral circulation that often leads to skin ulceration and diminished delivery of neutrophils to sites of microbial entry. Diabetes was prevalent in 70% of our patients with Stage V kidney function, but was not shown statistically to be a multivariate risk factor for vascular complications. Currently, there are no standardized guidelines for prevention of access-site infections. Some experts have recommended prophylactic use of antibiotics before ACD placement in high-risk patients, especially those with diabetes mellitus,29,30,34,35 however, no prospective clinical trials are available to assess the effectiveness of this strategy.

Arterial dissections were seen in 10 (1.6%), femoral artery thrombosis in 5 (0.8%), and device failure in 3 (0.5%) of our study patients. These complications have been published in prior studies, but have very low incidence rates and have shown no statistically significant differences between the types of devices used or between any other pre, intra- and post-procedural characteristics.36–38 This finding was similar to the results of our study.

Previous Studies

Multiple studies and three major meta-analyses have investigated the complication rates post-ACD use compared to mechanical compression.13,14,39–48 Despite the differences in the study population and design, the following conclusions can be drawn. Among patients undergoing diagnostic cardiac catheterization, there is a 0.5–1.7% rate of vascular complications and this risk is not consistently increased or decreased by ACD usage across all studies.18 Among patients undergoing interventional procedures, there is a 0.8–5.5% rate of heterogeneously defined vascular complications, and there are no clear data to indicate increased risk in the ACD groups, with the exception of the Vaso-Seal device.18 These studies identified older age, female gender, lower weight, smaller body surface area and renal failure to confer increased risk of vascular complications, but there are no studies specifically evaluating the role of kidney function on complication rates after ACD placement.37–40

Study limitations. This was a retrospective study and is subject to the limitations of this design. Also, the single-center experience limits the generalizability of these findings beyond a sample of patients with similar medical conditions and treated with similar medical protocols. Reporting biases because of variations in endpoint definitions may have influenced the apparent higher incidence of complications, but are unlikely to have biased the results. This study outlined in-hospital outcomes; however, the possibility of late complications cannot be extrapolated from our data. The more severe CKD groups had relatively small numbers of patients. Also, we have no comparison to nonfemoral artery access, but these patients with severe renal dysfunction would not be good candidates for radial or brachial artery access due to the need for hemodialysis fistula creation in their arms should they need permanent hemodialysis. Finally, there was no comparison to the manual compression group.

Conclusions

Our results suggest that patients with baseline CKD undergoing PTCP and ACD placement should be closely monitored for complications, especially infection and hematoma formation. The risk category can also be further stratified based on the degree of renal function with Stage IV CKD and above being at the highest risk for vascular complications. Our findings also reemphasize the importance of aseptic techniques and judicious use of prophylactic antibiotics. There is a need for large-scale randomized studies of these devices to definitively address safety concerns and also to further identify predictors of complications. Data on predictors of complications can play a pivotal role in prevention, early detection and proper management in these high-risk patients.

From the aDivision of Cardiology, St. Luke’s and Roosevelt Hospital Center, Columbia University College of Physicians & Surgeons, bLenox Hill Hospital, and cGood Samaritan Hospital, New York, New York.

The authors report no conflicts of interest regarding the content herein.

Manuscript submitted March 9, 2009, provisional acceptance given April 10, 2009, final version accepted September 25, 2009.

Address for correspondence: Mun K. Hong, MD, Director, Cardiac Catheterization Laboratory and Interventional Cardiology, St. Luke’s-Roosevelt Hospital Center, 1111 Amsterdam Avenue, New York, NY 10025. E-mail: MKHong@Chpnet.org