Angiographic Predictors of Vascular Complications Among Women Undergoing Cardiac Catheterization and Intervention

Angiographic Predictors of Vascular Complications Among Women Undergoing Cardiac Catheterization and Intervention
Angiographic Predictors of Vascular Complications Among Women Undergoing Cardiac Catheterization and Intervention
Angiographic Predictors of Vascular Complications Among Women Undergoing Cardiac Catheterization and Intervention
Angiographic Predictors of Vascular Complications Among Women Undergoing Cardiac Catheterization and Intervention
Angiographic Predictors of Vascular Complications Among Women Undergoing Cardiac Catheterization and Intervention
512 - 516

Bina Ahmed, MD, Stefan Lischke, MD, Leigh Ann Holterman BS, Faye Straight, RN, Harold L. Dauerman, MD

ABSTRACT: Background. Bleeding and vascular complications remain more common in women than men undergoing invasive cardiovascular procedures. We determined the role of femoral angiographic variables in risk-stratifying women for vascular complications. Methods. Between 2004–2009, all major bleeding and vascular complications among women undergoing diagnostic or interventional cardiovascular procedures were identified at a single center. Thirty consecutive female patients (major bleeding or vascular complication) were then age- and procedure-matched to 90 controls (no vascular complications). Quantitative femoral angiography was performed on all cases and controls. Results. Smaller minimum luminal diameter was a strong univariate predictor of vascular complications in women (odds ratio [OR] 0.65; 95% confidence interval [CI] 0.47–0.90; p = 0.009), while site of arteriotomy was not predictive of complications. The prognostic significance of smaller femoral lumen diameter was mildly attenuated after adjusting for the predictive factor of smaller patient body size, even after adjusting for the predictive factor of smaller patient body size. Finally, multivariable modeling suggests that utilization of vascular closure devices (OR 0.26, 95% CI 0.07–0.96; p = 0.04) may be protective in women. Conclusions. Women with smaller femoral arteries are at significantly higher risk for bleeding and vascular complications than women with larger femoral arteries. Risk stratification for bleeding complications among women should account for clinical, pharmacologic and femoral angiographic factors.

J INVASIVE CARDIOL 2010;22:512–516

Key words: women, bleeding complications, femoral artery
Over the last decade there have been significant improvements in bleeding and vascular complications (VC) among women undergoing cardiac catheterization.1 Several observational reports show that procedure-related improvements such as use of smaller sheaths, fluoroscopy-guided access to obtain “safe-zone” arteriotomy, and improved antithrombotic therapies have been associated with a 50% reduction in VC rates.2–5 However, when VC do occur, they continue to be associated with significant morbidity and mortality; procedure-related bleeding is among the strongest predictors of subsequent adverse cardiac events.6–9 A recent analysis of over 37,000 patients from the Northern New England Cardiovascular Database Registry continues to show a disturbing, and largely unexplained, adjusted 2.5 fold increase risk of VC in women compared to men undergoing percutaneous coronary intervention.1 One explanation for this persistent disparity in bleeding complications is a possible difference in femoral artery anatomy and access site suitability between the genders. While prior studies have demonstrated a difference in common femoral artery (CFA) diameters between men and women,10,11 the role of femoral angiographic variables in predicting bleeding and vascular complications specifically among women has not been studied. Accordingly, we sought to define angiographic predictors of bleeding and vascular complications in women undergoing cardiac catheterization and coronary intervention.


A case-control study was performed utilizing age and procedure (diagnostic only or percutaneous coronary intervention [PCI]) to match 30 women with VC (cases) to 90 women without VC (controls) and analyze the association of femoral angiographic variables with VC. Data were obtained from the University of Vermont’s interventional cardiology database for all women undergoing cardiac catheterization over a 5-year period (2004–2009). In addition to detailed data on demographics, medical history, laboratory, pharmacotherapy and procedure-related information, quantitative femoral angiographic analysis was performed by a single reviewer blinded to outcomes using an automated edge-detection software program (Philips Medical). VC was defined according to the Northern New England Cardiovascular Disease Study Group 1): hematoma > 6 cm, any access site related bleeding requiring transfusion or any injury requiring mechanical intervention. All VCs were confirmed using chart review to identify associated complications occurring during the index hospitalization. All patients underwent femoral angiography in an approximately 30-degree ipsilateral view. Quantitative femoral angiography was performed to determine the following variables: common femoral artery (CFA) reference vessel diameter; CFA minimum luminal diameter (MLD); CFA maximum stenosis; CFA length defined as distance from ostium of the inferior epigastric artery (IEA) to CFA bifurcation; anatomic site of arteriotomy (CFA, external iliac artery, profunda or superficial femoral artery). For patients with no visible femoral plaque, MLD and reference vessel diameter are the same. Safe-zone arteriotomy was defined angiographically as sheath insertion between the inferior border of the inferior epigastric artery and above the CFA bifurcation (Figure 1).12,13 Safe-zone length was defined as the distance between inferior border of IEA and the CFA bifurcation. High bifurcation was defined as CFA bifurcation above the inferior border of the femoral head.

Statistical analysis. Continuous variables are expressed as mean ± standard deviation and frequencies are expressed as percentages. Due to the multiple case-control matching ratio of 1 to 3, univariate and multivariate comparisons between cases and controls used conditional logistic regression methods. These univariate and multivariate methods were used to determine univariate predictors of VC as well as predictors for VC after adjusting for significant clinical differences between the cases and controls. To assess whether CFA minimum luminal diameter was an independent predictor of VC among women, multivariate analysis was performed incorporating clinical variables known to be associated with femoral dimensions (body mass index [BMI],11 as well as potentially confounding angiographic variables shown to be univariate predictors (high bifurcation anatomy). Given the prior description of safezone arteriotomy as protective against VC,12 safe-zone arteriotomy was also controlled for in this model. The multivariate analysis was then repeated incorporating vascular closure devices in order to understand the interaction between vessel size, closure devices and complications among women. All analyses were generated using SAS software, version 9.2 for Windows (SAS, Inc., Chicago, Illinois).


The mean age of the entire group was 66.5 ± 12.1 years. Approximately 30% of all patients had a high bifurcation of the CFA, and the mean CFA length was 37.86 ± 9.36 mm. VC in cases were clinically severe: 50% of cases had retroperitoneal bleeds, 50% required blood transfusion and 43% required mechanical intervention due to bleeding/vascular complications. When comparing cases and controls, women with VC had a significantly lower BMI compared to women without a VC (26.6 ± 6.7 vs. 31.2 ± 6.8; p = 0.002) and were more likely to present with ST-elevation myocardial infarction (STEMI) (24% vs. 8%; p = 0.05) (Table 1).

Creatinine clearance tended to be lower in women with complications. Women with VC were also more likely to receive glycoprotein IIb/IIIa inhibitors during PCI, whereas women without VC were more likely to receive bivalirudin during PCI. In addition, vascular closure devices were used more frequently in women without VC compared to women with VC (92% vs. 70%; p < 0.001).

Quantitative femoral angiography showed that women with VC had a smaller CFA reference vessel diameter (5.4 ± 1.2 mm vs. 5.9 ± 1.1 mm; p < 0.09) and MLD (4.6 ± 1.4 mm vs. 5.4 ± 1.3 mm; p = 0.009) with greater disease burden (8.1 ± 10.5 vs. 16.4 ± 14.5% stenosis; p = 0.001) than female controls (Table 2). Total femoral artery length was similar between the groups, although women with VCs were more likely to have a high bifurcation (47% vs. 29%; p = 0.05). On conditional logistic regression, BMI (OR 0.88; 95% CI 0.81–0.96; p = 0.002), CFA MLD (OR 0.65; 95% CI 0.47–0.90; p = 0.009) and lack of VCD use (OR 0.20; 95% CI 0.06–0.60; p = 0.005) all were potent predictors of VC, while CFA reference vessel diameter trended toward significance (OR 0.73; 95% CI 0.51–1.05; p = 0.09). When the entire patient group was stratified in quartiles based on CFA MLD, women in the lowest quartile were three times more likely to have a VC compared to women in the largest quartile (p = 0.03 for trend) (Figure 2).

The overall distribution of arteriotomy site location was similar between cases and controls: women with VCs were just as likely to have a safe-zone arteriotomy as women without VCs (77% vs. 77%; p = NS). Furthermore, among female cases specifically identified for retroperitoneal bleeding, three-quarters of those patients had a documented safe-zone arteriotomy.

In order to understand the potential interaction of patient size (BMI) on femoral angiographic variables, we performed a multivariate conditional logistic regression to analyze the influence of MLD on VC while controlling for BMI and two other potential angiographic predictors of VC (high bifurcation and lack of safe-zone arteriotomy). CFA MLD impact on risk of VC was mildly attenuated in this model (OR 0.69, C.I. 0.45–1.02; p = 0.06), while higher BMI remained a protective factor (OR. 0.90, 95% C.I 0.83–0.98; p = 0.03). High bifurcation and absence of a safe-zone arteriotomy were not independent predictors of VC in this model (Table 3). Finally, we explored the potential interaction of closure devices on VC among women. In a multivariate analysis controlling for both BMI and potential angiographic risk factors (MLD, high bifurcation and non-safe-zone arteriotomy), closure device utilization remained highly protective (OR 0.26, 95% CI 0.07–0.96; p = 0.04). In this model, higher BMI remained protective (0.91, 95% CI 0.84–0.99; p = 0.03), MLD impact was attenuated (OR 0.72, 95% CI 0.47–1.1; p = 0.17) and once again, safe-zone arteriotomy did not prove protective against VC (p = 0.44).


Our case-control study confirms that among women undergoing cardiac catheterization or coronary intervention, common femoral artery vessel size is inversely associated with a significant increased risk of VC: women with the smallest femoral artery MLD are three times more likely to have a VC than those with the largest femoral arteries. On the other hand, we found no clear relationship between site of arteriotomy and a propensity for VC in women.

To our knowledge, this is the first study to specifically evaluate the role of femoral anatomy as it relates to the incidence of bleeding and VCs among women. There is ample evidence of associations between clinical factors (gender, BMI, renal failure, diabetes, cardiogenic shock, peripheral vascular disease)4,5,12,14 and vascular/bleeding complications. Whether femoral artery anatomy contributes to the higher incidence of vascular complications in women remains understudied. With changes in contemporary practice resulting in overall reduction in vascular complication rates, we expected to see a narrowing of the gender gap; however this has not been observed and women remain at an increased risk of VC despite rigorous adjustment for clinical co-morbidities.1

Post new comment

  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.

More information about formatting options

Enter the characters shown in the image.