Featured Article

Introduction. Systemic hypotension has been observed for up to 72 hours in response to stimulation of the carotid baroreceptor by carotid angioplasty and stenting (CAS). This study sought to identify risk factors and cardiac outcomes for post procedural hypotension after carotid artery angioplasty and stenting. Methods. Between 2003 and 2005, 143 patients (87 men, mean age 75) with high-grade carotid artery stenosis (mean, 87.3%) were treated with CAS. Atropine and appropriate intravenous crystalloid solution were administered during CAS. Hypotension was treated with vasopressors (dopamine, phenylephrine or norepinphrine). Patients were stratified into three groups based on hypotension (systolic blood pressure < 90) requiring vasopressors for 0 hours ( no vasopressors ), ? 24 hours (short duration) and > 24 hours (prolonged duration). Risk factors for hypotension requiring vasopressors were analyzed by univariate and multivariate logistic regression analysis. Results. Post procedural (not including intraprocedure vasopressors) hypotension requiring vasopressor treatment was seen in 16/143 (11%) of patients, with 6/143 (4%) requiring vasopressor support for more than 24 hours. Mean duration of vasopressor administration was 23 hours (range 6–72 hours). By univariate analysis, a history of a previous MI (p = .02), or use of the PercuSurge occlusion balloon (p = .05) were both associated with increased incidence of short duration vasopressors (? 24 hrs) use, while female sex (p = .03) and age > 80 (p = .02) were associated with prolonged vasopressor requirement (> 24 hours). On multivariate analysis, a history of MI (OR 4.2, 95% CI 1.1–16.3, p = .04) remained an independent predictor of short-duration vasopressors, and female sex (OR 10.9, 95% CI 1.2–100.4, p = .04) and age > 80 years old (OR 13.8, 95% CI 1.5–127.2, p = .02) remained predictors of prolonged vasopressor use. The incidence of periprocedural MI, arrhythmias or cardiac heart failure did not differ between those patients who did not receive vasopressors (5/127) and those that received vasopressors for a short duration (? 24 hours = 1/10, p = NS) or prolonged duration (> 24 hours = 0/6, p = NS). These findings suggest that post-operative hypotension requiring vasopressors is not caused by and does not induce peri-procedural cardiac complications. Conclusion. Prolonged hypotension requiring vasopressor support is rare after CAS, with higher incidences in older females. Prolonged hypotension requiring vasopressors is not associated with an increased incidence of periprocedural cardiac complications.

Introduction
In medically high-risk patients, carotid angioplasty and stenting (CAS) is increasingly being performed as an alternative to carotid endarterectomy (CEA) for the treatment of carotid artery stenosis. Hemodynamic instability following CEA has been observed in 12–54% of cases.^1,2 This post-operative hemodynamic instability following CEA is associated with increased rates of neurological and cardiac complications.^2–5 Persistent hypotension following CAS has been reported in 7–34% of cases, but the associated risk factors and potential clinical sequalae remain incompletely characterized.^6,7 This study sought to identify risk factors and cardiac outcomes associated with post-procedural hypotension requiring vasopressors after CAS, and to stratify patients based on the duration of vasopressor use.

Methods
Demographics. From January 2003–January 2005, a total of 143 patients (87 men, mean age 74.5) with carotid artery stenosis were treated with CAS at one institution. Data were entered prospectively into a computerized registry under the approval of the Institutional Review Board. Analysis of these data were preformed retrospectively. Preoperative patient demographic data are listed in Table 1. Recorded procedural details included heart rate, blood pressure, medications given, degree of stenosis, size of native vessel, side of procedure, stent type and size, protection device used and angioplasty balloon size. Postoperative data collected included heart rate, blood pressure, medications, cardiac enzymes, EKG changes, morbidity, mortality and length of stay.

Hypotension was defined as a systolic blood pressure < 90 mm Hg. Vasopressor treatment was initiated in the absence of bleeding, hypovolemia or acute heart failure. Bradycardia was defined as heart rate < 40 bpm. Patients were stratified into three groups, based on hypotension requiring no vasopressor support (0 hrs), ? 24 hours of vasopressors or > 24 hours of vasopressors. Perioperative cardiac complications were defined as myocardial infarction (MI) as determined by EKG changes or elevations of cardiac enzymes (troponin), onset of new arrhythmias or congestive heart failure.

The degree of stenosis was determined by angiographic measurement criteria from the North American Symptomatic Carotid Endarterectomy Trial with a mean stenosis of 87.3% (range 60–99%).⁸ In this study group, 40 (28%) of patients experienced preoperative neurological symptoms, and 103 (72%) were asymptomatic. All patients were considered to be high risk for periprocedural complications for CEA. Forty-eight patients were over 80 years old. Twenty-one patients presented with restenosis after carotid endarterectomy, with one additional patient having restenosis after CAS. Six patients had carotid stenosis after prior neck surgery with associated irradiation therapy. Occlusion of the contralateral carotid artery was present in 17 patients. One hundred and sixteen patients were at increased risk for cardiac complications based on the Eagle criteria (moderate risk = 108 patients, high risk = 18 patients).⁹ Eighteen patients had severe chronic obstructive pulmonary disease.

Carotid angioplasty and stenting technique. All patients underwent CAS under local anesthesia without sedation. Procedures were performed by vascular surgeons in the operating room angiography suite using a fixed imaging system (Siemens AG, Munich, Germany). Femoral access was utilized in all cases. Cerebral protection devices were used in 140/143 cases, excluding 3 patients who had recent restenosis after CEA (n = 2) or CAS (n = 1). Microporous filters were used in the majority of cases (n = 105) and included EPI FilterWire (Boston Scientific, Maple Grove, Minnesota), Accunet (Guidant, Santa Clara, California) and Angioguard (Cordis, Miami, Florida). Temporary balloon occlusion with the PercuSurge GuardWire (Medtronic, Santa Rosa, California) was employed in 35 cases. Once the protection device was deployed, prophylactic atropine (0.5–1.0 mg) was administered.¹⁰ After the resultant increase in heart rate was observed, the lesion was predilated with a 4 x 40 mm monorail angioplasty balloon (Long VIVA, Boston Scientific). A self-expanding stent was then deployed across the lesion. Types of stents used included Wallstent (Boston Scientific), Acculink (Guidant), Precise (Cordis), AVE stent (Medtronic) and NexStent (Endotex, Cupertino, California). Post-dilatation with a monorail balloon 5–6 x 20 mm (Gazelle, Boston Scientific) was performed. Nitroglycerin (100 mcg) was administered for treatment of spasm of the distal internal carotid artery in 6 cases and was not associated with hypotension or vasopressor use.

Hemodynamic monitoring. Heart rate and rhythm were monitored by continuous EKG throughout the procedure and during the post-procedural period. Euvolemia was monitored in the periprocedural period by assessing urine output, tachycardia and blood pressure. All patients received intravenous fluid bolus prior to initiating the angioplasty procedure. Antihypertensive medication was not withheld prior to the procedure. In particular, beta-blockers were continued in the preoperative period. An automated cuff inflation sphygmomanometer was used on the upper arm to measure blood pressure at 5-minute intervals throughout the procedure and during the post-procedural period. Continuous blood pressure monitoring was performed during the procedure via the femoral sheath. When post-operative hemodynamic considerations such as vasopressor administration required continuous arterial pressure monitoring, a radial arterial line was placed. After the procedure, patients were admitted for a minimum of 24 hours, during which time heart rate, respiratory rate and blood pressure continued to be monitored. Electrocardiograms were routinely performed in all patients post-operatively, and compared with the preoperative electrocardiogram. Cardiac isoenzymes were obtained for all patients with a history of coronary artery disease (CAD), in all patients who experienced hemodynamic instability or symptoms suggestive of coronary ischemia, and in accordance with specific study protocols.

Pharmacologic measures. All the patients received clopidogrel (Sanofi-Synthelabo Pharmaceuticals, New York, New York) prior to the initiation of the CAS procedure and were maintained on clopidogrel for at least 30 days postoperatively. Clopidogrel was initiated 5 days prior to CAS, or a loading dose of 300 mg was given the day prior to CAS. Anticoagulation with heparin sodium was maintained throughout the procedure with a target activated clotting time of 250–350 seconds. Atropine (0.5–1.5 mg) was given for the treatment of reflex bradycardia after balloon dilatation of the carotid baroreceptor. For the initial 14 patients treated in the study, atropine was administered selectively for the development of bradycardia with a heart rate < 40 bpm or greater than 50% change in heart rate. Subsequently, patients with primary carotid artery stenoses received a prophylactic dose prior to balloon dilatation. Intravenous vasoactive agents (dopamine, phenylephrine or norepinephrine) were used for protracted hypotension.

Statistical analysis. Two separate analyses were done to determine predictors of post-operative hypotension requiring vasopressor support and cardiac morbidity. First, patients with post-procedural hypotension requiring short duration vasopressor support (? 24 hrs) were compared to those requiring no vasopressors (0 hrs, including those without post-procedural hypotension. Second, patients with post-procedural hypotension requiring prolonged duration vasopressor support (> 24 hrs) were compared with those receiving no vasopressors (0 hrs) or only short-duration vasopressors (? 24 hrs). Univariate testing was done using Fisher’s exact test for analysis of dichotomous data and t-tests for continuous data analysis. Continuous data are expressed as mean ± standard deviation. Significance was assumed at p ? .05 and all tests were two-tailed. A multivariate logistic regression model with odds ratios (OR) and 95% confidence intervals (CI) was developed based on the results of the univariate analysis, with inclusion variables having a value of p ? .05. Dependent variables were either short-duration vasopressor support or prolonged-duration vasopressor support. Independent variables used in each model were selected from those listed in Tables 2 and 5. Cardiac outcomes were analyzed with Fisher’s exact test. Data were managed and analyzed by SPSS statistical software (SPSS 10.1, SPSS Inc., Chicago, Illinois).

Results
CAS was performed successfully in all 143 patients. Peri- or post-procedural hypotension was seen in 27 (19%) patients. Vasopressor support for post-procedural hypotension was required in 16/143 patients (11%), with 10/143 (7%) requiring ? 24 hrs and 6/143 (4%) requiring more than 24 hours of vasopressors. Peri- and post-procedural bradycardia was observed in 16 (11%) patients, and was more common in those requiring vasopressor support both short (7/10, p < .001) and prolonged (3/6, p < .001). No patients with post-procedural bradycardia required additional atropine administration or insertion of a temporary pacemaker.
Predictors of short duration vasopressors (? 24 hours). Ten patients experienced hypotension, requiring vasopressors for ? 24 hours (mean 13.7 ± 5.4, range 6–22 hrs). The baseline demographics and characteristics of the patients are shown in Table 2. These 10 patients (6 males:4 females) had a mean age of 70.0 ± 10.1, with 5/10 having preoperative neurological symptoms and a mean stenosis of 91.1%, with none of the factors differing significantly from the group no vasopressors [n = 127, p = not significant (NS)]. Univariate analysis of demographic factors demonstrated that short-duration vasopressor support for hypotension was significantly associated with a history of MI (no vasopressors: 29/127 = 23% versus ? 24 hrs of vasopressors: 6/10 = 60%, p = .02). Patients requiring short-duration vasopressors had elevated rates of other co-morbidities without reaching statistical significance, including aortic stenosis, CAD, recent coronary artery revascularization and diabetes mellitus. No significant difference was observed in history of hypertension, hypercholesterolemia, renal insufficiency or smoking history. The incidence of high-risk factors was not significantly different between the groups, including age > 80, history of prior CEA or CAS, history of neck irradiation, contralateral occlusion, moderate (1–2 Eagle criteria) or high (> 3 Eagle criteria) cardiac risk or severe chronic obstructive pulmonary disease (COPD) (p = NS).

By univariate analysis, intra-procedural factors (Table 3) significantly associated with an increased incidence of post-procedural hypotension requiring vasopressors for ? 24 hours included use of the PercuSurge occlusion balloon (no vasopressors: 27/127 = 21% versus ? 24 hrs of vasopressors: 5/10 = 50%, p = .05), intra-procedural hypotension (no vasopressors: 11/127 = 9% versus ? 24 hrs of vasopressors: 8/10 = 80%, p < .001) or intra-procedural use of vasopressors (no post-procedural vasopressors: 10/127= 8% versus ? 24 hrs of vasopressors: 8/10 = 80%, p < .001). Intra-procedural bradycardia was not associated with an increased incidence of post-procedural hypotension (p = NS, Table 3).

Logistic regression multivariate analysis demonstrated that a history of prior MI (OR: 4.2, 95% CI = 1.1–16.3, p = .04) was independently associated with a need for post-procedural vasopressors for ? 24 hours (Table 4). On multivariate analysis, the use of the PercuSurge occlusion balloon was not significantly associated with an increased likelihood of post-procedural vasopressors for 24 hours (OR 2.8, 95% CI = 0.72–10.9, p = 0.14).
Predictors of hypotension requiring prolonged vasopressor support (> 24 hours). Six patients experienced hypotension requiring vasopressors for greater than 24 hours (mean 38.3 ± 18.5 hrs, range 26–72 hrs). The baseline demographics and characteristics of the patients are shown in Table 5. These six patients (> 24 hours of vasopressors) were compared with patients who received 0–24 hours of vasopressor support (included no vasopressors and those that required vasopressors for ? 24 hours). Five out of six patients requiring prolonged vasopressors (> 24 hrs) were female, with a significantly higher mean age of 80.8 ± 5.1 years (p = .02), compared with those receiving no vasopressors or ? 24 hours of vasopressor administration (n = 137). Univariate analysis of demographic factors demonstrated that patients requiring prolonged vasopressors had rates of co-morbidities, including history of hypertension, MI, pre-operative neurological symptoms and diabetes mellitus, similar to patients who required no or ? 24 hours of vasopressor administration. On univariate analysis, the only high-risk pre-procedural factor significantly associated with risk of hypotension requiring prolonged vasopressors was age > 80 (0–24 hrs of vasopressors: 43/137 = 31% versus > 24 hrs of vasopressors: 5/6 = 83%, p = .02). Otherwise, this group of patients had similar rates of prior ipsilateral CEA or CAS, history of neck irradiation, contralateral occlusion, and moderate (1–2 Eagle criteria) or high (> 3 Eagle criteria) cardiac risk or severe COPD (p = NS).

Intra-procedural factors (Table 6) predictive of post-procedural hypotension requiring prolonged vasopressors were intra-procedural hypotension (0–24 hrs of vasopressors: 19/137 = 13% versus > 24 hrs of vasopressors: 6/6 = 100%, p < .001) and intra-procedural bradycardia, (0–24 hrs of vasopressors: 15/137 = 11% versus > 24 hrs of vasopressors: 3/6 = 50%, p = .03). The use of the PercuSurge balloon and intra-procedural vasopressors were not significantly more likely to predict post-procedural hypotension requiring prolonged vasopressors.

Multiple logistic regression multivariate analysis demonstrated that age > 80 (OR 13.8, 95% CI= 1.5–127.2, p = .02), and female sex (OR 10.9, 95% CI = 1.2–100.9, p = .04) were independently associated with post-procedural hypotension requiring prolonged vasopressors (> 24 hours, Table 7).

Cardiac morbidity. Cardiac morbidity was defined as a periprocedural MI, arrhythmias or congestive heart failure (Table 8). Three patients (2.1%) had a periprocedural MI as determined by analysis of troponin levels, which was not more common in patients receiving short-duration vasopressors (? 24 hours = 1/10, p = NS) or prolonged vasopressors (> 24 hours = 0/6) compared with patients not requiring vasopressors (1/127, p = NS). Three patients had a cardiac arrhythmia requiring treatment, which also did not significantly differ between groups (no vasopressors =1/127; ? 24 hours of vasopressors =1/10; > 24 hours vasopressors = 0/6, p = NS). The one patient that had an acute episode of congestive heart failure did not require any vasopressors either prior to or after the episode. The length of stay did not differ significantly between groups [no vasopressors: 1.4 ± 1.3 days (d), 1–24 hrs of vasopressors: 3.6 ± 5.5 d, > 24 hrs of vasopressors: 1.8 ± 0.4 d, p = NS].

Discussion
The hemodynamic instability that occurs both during and after CAS is mediated through the baroreceptors, which are located in the adventia at the carotid bifurcation.¹¹ These baroreceptors are stimulated by the stretch of balloon angioplasty, initiating a reflex arc to the nucleus tractus solitarius via the sinus nerve of Hering, a branch of the glossopharyngeal nerve. This in turn causes a reciprocal change in peripheral sympathetic and vagal neural activity, resulting in a fall in blood pressure and bradycardia. Stent placement may promote persistent stimulation of these baroreceptors.

In this study, two distinct clinical entities appear to be present after CAS. The first group involves hypotension requiring vasopressor administration over a relatively acute period. Patients who develop this experience rapid resolution within 24 hours, and frequently within 6 hours. These patients who required a short duration of vasopressors had elevated rates of history of previous MI. This may be explained in part by data suggesting increased carotid artery baroreceptor sensitivity in patients with CAD.^12,13 Previous studies have shown that patients with cardiac risk factors, such as CAD,6 unstable angina¹⁴ and previous MI,¹⁵ are more likely to have post-procedural hypotension after CAS. However, these studies have great variability in the duration of the hypotension, and do not specifically examine predictors of vasopressor use.

The second clinical group manifested a more protracted course of hypotension necessitating vasopressor support. This entity appears to be unrelated to the patients’ underlying cardiac conditions. Rather it may represent a prolonged reflexive response to which elderly women exhibit greater susceptibility. Previous studies have found age to be a risk factor for hemodynamic instability during CAS, which may be related to dysfunction of baroreceptors.⁶ In addition, physiological changes of older age, including low blood pressure and age-related ventricular dysfunction, may increase the susceptibility to hypotension.⁶ Early reports about CAS demonstrated that older age was a risk factor for post-operative complications.^16–19 In the current study, patients who experienced prolonged hypotension after CAS did not exhibit any increase in cardiac or other complications. This implies that while this patient cohort may be more susceptible to hypotension, the development of hypotension does not result in adverse consequences.

Intra-procedural hypotension strongly predicted the occurrence of post-procedural hypotension both in this current study and in previous reports.^14,15 In a smaller study involving 51 patients, Qureshi et al. found that intra-procedural hypotension was the strongest predictor of post-procedural hypotension. Another study found that balloon expandable stents with larger final balloon diameters rather than self-expanding stents were associated with post-CAS hypotension.¹⁴

Limitations of this study include the relatively small sample size, variation in types of vasopressors used and small number of adverse outcomes, which reduces the number of factors that can be integrated into a multivariate model. In addition, since the majority of patients presented with significant co-morbidities, it may make distinguishing these at-risk factors more difficult. However, this study reemphasizes the importance of close hemodynamic monitoring after CAS. This may also explain why significant differences were not present in the patient characteristics among the three groups. In addition, alternative strategies to allow patients to be treated with vasopressors for less than 24 hours can be considered. Of potential use would be the discontinuation of antihypertensive medications prior to carotid angioplasty. In particular, this might be considered in elderly women who do not have significant CAD.

In conclusion, prolonged hypotension requiring vasopressor support is rare after CAS, with higher incidences in older females. Prolonged hypotension requiring vasopressors is not associated with an increased incidence of periprocedural cardiac complications. In contrast, hypotension that is more limited in duration occurs more commonly in patients with CAD who had a prior MI. Continued examination of these phenomena will help further characterize which patients can be anticipated to have periprocedural hemodynamic variability, and therefore perhaps require a more intensive level of observation.