| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Both authors: Department of Radiology, Beth Israel Deaconess Medical Center, 330 Brookline Ave., Boston, MA 02215.
Received August 3, 2006;
accepted after revision November 21, 2006.
Address correspondence to R. G. Sheiman.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. A history of peripheral vascular disease, a
prior cerebrovascular event, smoking, diabetes, hypertension, cervical carotid
disease, left main coronary disease, and patient sex were criteria
prospectively gathered for 295 consecutive patients undergoing screening
carotid sonography before elective coronary artery bypass surgery. Logistic
regression modeling was used to determine if any single criterion or
combination of criteria could be applied to decrease the number of screening
examinations without sacrificing detection of significant (
50%
cross-sectional narrowing) carotid stenosis.
RESULTS. Smoking, diabetes, hypertension, a previous cerebrovascular event, peripheral vascular disease, left main coronary artery disease, and a history of cervical carotid disease were associated with significant carotid disease (chi-square test) in our subject population. Logistic modeling showed that the probability of detecting significant carotid disease increases 2.98 times for each additional selection criterion present. Possessing at least one selection criterion would still yield 100% examination sensitivity while increasing specificity to 30.0%.
CONCLUSION. Selection criteria should be applied when choosing patients for carotid sonographic screening before elective coronary artery bypass surgery. This approach would decrease the number of noncontributory examinations but would have little effect on the detection of significant carotid stenosis in this target population.
Keywords: cardiovascular disease • cardiovascular imaging • Doppler sonography • screening • vascular imaging
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
50% cross-sectional narrowing) as a significant risk factor for
the development of a cerebrovascular event after CABGS
[1-7].
These studies, however, are partially undermined by the actual reported
prevalence of significant carotid disease in such patients (17-22%
[6,
8-10])
and the overall incidence of cerebrovascular events after CABGS (1-6%,
depending on its definition [3,
5,
6,
8]). Hence, significant carotid
disease should, at most, account for only approximately 1% of all
cerebrovascular events resulting from CABGS. Superimposed on this are multiple
studies using multivariate analysis that actually dispute any relationship
between carotid disease and a cerebrovascular event after CABGS
[5,
11]. Despite this controversy, documentation of carotid disease before CABGS is believed by many to be of paramount importance. Thus, carotid sonographic screening (duplex and color Doppler sonography) is readily performed before elective bypass surgery at many institutions to identify such patients, although the clinical approach taken after their identification remains in contention (initial endarterectomy, combined endarterectomy and bypass, close postbypass monitoring with later endarterectomy) [8, 9-14].
Overall, the use of coronary artery disease as a sole selection criterion for performing preoperative carotid sonographic screening is flawed by its low yield for identifying significant disease, the lack of a standard clinical approach taken to patients with a positive screening examination, and the prevalence of patients who actually have a postoperative cerebrovascular event. A logical first approach to this controversy would be to try to target preoperative sonographic screening to only those at high risk for significant carotid disease. Therefore, we sought to identify selection criteria for performing carotid sonographic screening before elective cardiac bypass in hopes of decreasing unnecessary examinations without sacrificing the detection of significant carotid disease. Our purpose was not to identify patients at high risk for a cerebrovascular event after CABGS or to define the incidence of carotid disease in patients undergoing CABGS because these issues have already been addressed in the literature.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Imaging
Examinations were performed by one of three registered vascular
technologists, all with a minimum of 15 years of experience, on an ATL 5000
scanner (ATL) using a 7.5- or 10-MHz transducer. Our standardized examination
included duplex and color Doppler imaging of the common, internal, and
external carotid arteries (CCA, ICA, and ECA, respectively) in the sagittal
and transverse planes and determination of peak systolic and diastolic
velocities of the proximal and distal CCA; proximal, mid, and distal ICA; and
proximal ECA. Vertebral artery interrogation for peak systolic velocity and
flow direction was also performed. The angle of interrogation was less than or
equal to 60°, with the pulse repetition frequency, wall filter, and
Doppler gain optimized for each patient. All studies were interpreted by one
of two vascular radiologists with a minimum of 7 years of experience. Criteria
for grading stenosis were those currently recommended by the Society of
Radiologists in Ultrasound
[15].
Criteria
Our goal was to identify preoperative selection criteria that could be
applied to prospectively identify the subpopulation who could potentially
benefit from carotid sonography before elective cardiac bypass. Potential
criteria proposed were chosen on the basis of risk factors already shown in
the literature to be independently and significantly associated with either
carotid disease or a cerebrovascular event after cardiac bypass surgery. These
included history of a prior cerebrovascular event (defined as a
cerebrovascular accident or transient ischemic attack), female sex, diabetes
[3,
9,
11], left main coronary artery
disease [8,
9], peripheral vascular disease
[6,
9,
16], hypertension
[3,
16], and smoking
[9,
10]. For the sake of
completeness, a history of cervical carotid disease was also included.
Proposed selection criteria for carotid sonography for each patient were obtained by electronic chart review immediately before the examination. For a patient to have a history of a prior cerebrovascular event, documentation was required via formal neurologic consultation or confirmation via MRI or CT. Peripheral vascular disease was confirmed via prior lower extremity surgical bypass, imaging (MR angiography, CT angiography, or conventional angiography), or noninvasive arterial testing at our institution. Prior carotid disease required a history of an endarterectomy or disease documented in the medical record by a patient's primary care physician. All selection criteria for each patient were confirmed via a short (< 3 minutes) medical history obtained just before beginning the examination. When a discrepancy existed between chart review and patient history regarding the presence of a selection criterion, final determination was based on patient history. No attempt was made to correlate the extent of any criteria (e.g., pack-year smoking history) with the extent of carotid disease. Data acquisition consisted of the presence or absence of these parameters only and not the extent to which they existed.
Statistical Analysis
For the sake of analysis and as has been standard in the literature focused
on carotid disease as a risk factor for a cerebrovascular event after CABGS,
significant carotid stenosis was defined as at least a 50% diameter reduction
of either cervical ICA. Univariate analysis (chi-square test) was applied to
determine which of our proposed selection criteria were statistically
associated (p < 0.05) with sonographically confirmed significant
carotid stenosis in our subject population.
Multivariate logistic regression modeling was then performed in two stages. First, to identify which of the selection criteria found to be significant by univariate analysis would be independently associated with predicting significant carotid disease—that is, maintained after multivariate analysis. After this, each independently significant criterion identified by multivariate analysis was assigned a value of 1.0 so that a score could be assigned to any potential patient, representing the sum of the number of independent selection criteria possessed by that patient. Multivariate logistic regression modeling using only this score as the independent variable for any patient (rather than a patient's possessing any specific variables [criteria]) then yielded the probability of each score for predicting significant disease on a screening carotid examination.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
The average number of selection criteria for the 72 patients with carotid disease was 3.1, compared with 1.5 for those without disease. Forty-nine (68%) of 72 patients with disease had at least three selection criteria compared with only 51 (23%) of 223 without disease (p < 0.001, chi-square test). All 72 patients with significant disease possessed at least one criterion.
Multivariate logistic regression modeling maintained all seven criteria that were found significant as independently associated with carotid disease. Thus, using a scoring system of 0-7, our model indicated that the odds of obtaining a positive carotid sonography examination are 2.98 times greater for a patient with a score of N + 1 selection criteria compared with a patient with a score of N criteria. Table 2 shows the model's sensitivity, specificity, and likelihood ratio (likelihood of a positive sonographic examination for a given score divided by the likelihood of a negative sonographic examination for a score of 0) for predicting significant carotid disease on a screening examination as a function of score when applied to our population. A score of 0 or above—that is, performing a preoperative carotid sonography on all patients undergoing an elective cardiac bypass—has 100% sensitivity but 0% specificity for significant carotid disease. A score of 1.0 would maintain 100% sensitivity and increase specificity to 30%—or, in other words, detect all 72 of our patients with carotid disease while eliminating 67 of 295 examinations. Based on the cost of an elective carotid duplex examination at our institution, this would have saved $46,766 over the period of our study. Using a threshold score of two selection criteria would overlook seven of 72 patients with carotid disease while eliminating 119 of 295 examinations.
|
Although this was not the focus of our study, based on the discharge summary of each subject, 287 of our 295 subjects underwent coronary bypass surgery, with none experiencing a postoperative or perioperative cerebrovascular event. The remaining eight patients underwent medical management rather than elective bypass surgery because of their possessing a high-grade carotid stenosis (all had at least one carotid stenosis > 80%) and additional comorbidities. Of these eight, one underwent an uneventful carotid endarterectomy, six underwent unilateral carotid stenting, and one was lost to follow-up.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In published studies in which carotid disease was identified as a risk factor for a cerebrovascular event in the immediate post-CABGS period, the actual incidence attributed directly to carotid disease was extremely low. In the work of Schwartz et al. [6], 62.5% (5/8) of postoperative strokes were believed to be related to ipsilateral carotid disease. However, a different but conservative interpretation of their data is that screening sonography was of no immediate value in 70% of patients, and that only 0.86% (5/582) of patients screened had a stroke believed to be related to carotid disease. In another publication [3], 16,528 patients underwent cardiac bypass, with 333 (2.0%) experiencing a major neurologic deficit, but in only 16 (0.1%) patients was the deficit attributed directly to ipsilateral carotid stenosis. Hence, when one looks at the overall incidence of postoperative cerebrovascular events believed to be attributed directly to carotid disease in these studies implicating stenosis, the need to establish criteria to identify a subset of patients to undergo preoperative sonography becomes apparent.
Based on our results, carrying out screening carotid sonography on all individuals before elective CABGS is not required. Both our data and logistic regression modeling clearly show that this results in an extremely low specificity (0%) for such an examination.
To date, no selection criteria have been definitively proposed to identify patients for carotid sonographic screening before elective CABGS. The criteria we used clearly show that tailored selection of patients who require preoperative assessment can be performed to achieve an acceptable sensitivity and specificity for predicting significant carotid disease. Simply requiring one criterion to send a patient to undergo screening could have allowed us to identify all of our patients with significant carotid disease while increasing examination specificity from 0% to 30%. Note that, because each significant criterion was so strongly associated with carotid disease, in practice no single criterion needs to be considered more important than another, making a scoring system similar to the one used here a valid and simple approach to screening.
In support of the utility in applying selection criteria before performing a screening carotid examination, many investigators have found a negligible occurrence of post-CABGS cerebrovascular events in patients with no history of a cerebrovascular accident or transient ischemic attack. In the work by one group of investigators, of 1,279 subjects undergoing isolated coronary artery bypass, postoperative stroke attributed to carotid disease occurred in only two subjects, both of whom had a history of a prior cerebrovascular event [9]. Gerraty et al. [10] performed preoperative duplex sonography on 358 patients undergoing coronary bypass (n = 213) or peripheral vascular surgery (n = 145). A history of cerebral ischemia was documented in 49 patients and 309 were asymptomatic in this regard. Postoperative cerebral injury occurred only in symptomatic patients. Of note, the asymptomatic group also included 28 patients with carotid stenosis of 80% or greater. Those authors concluded that the risk of postoperative stroke in a patient with no history of any form of cerebral ischemia is low. Such a conclusion has also been drawn by others [2, 12, 13, 17] and indirectly seems to discourage mandatory carotid sonography before CABGS.
The findings of our prospective study are also supported by and logically extrapolated from the numerous publications that have identified significant risk factors for carotid disease in patients undergoing CABGS. For example, Salasidis et al. [16] found a history of peripheral vascular disease in 81 (20.9%) of 387 patients scheduled for nonemergent CABGS; and among these 81 subjects, 21 had significant carotid stenosis. D'Agastino et al. [9] had a similar prevalence of peripheral vascular disease in their population undergoing first-time CABGS (342/1,835 [18.6%]). Both investigators found peripheral vascular disease to be significantly associated with carotid stenosis in patients scheduled for cardiac bypass. Hence, using peripheral vascular disease as a selection criterion for undergoing screening preoperative carotid sonography logically follows. Despite the existence of such literature, formal recommendations and methods for selective carotid screening are not put forth.
Recommending performance of screening carotid sonography for patients with a history of carotid disease seems self-fulfilling and also needs some explanation. In our experience, the presence of carotid disease, whether communicated by the patient or noted in the medical record, can range from simple plaque with no associated stenosis to significant disease. Because of the variability of what is perceived as carotid disease, we prospectively believed that a history of carotid disease was a necessary parameter to evaluate. Although we found this parameter/criterion to be a significant predictor of carotid disease that would seem obvious, indeed, 15% (8/54) of our patients with a "history of carotid disease" as we defined it, had only hemodynamically insignificant plaque found on their preoperative carotid duplex examination.
Our study has several limitations. First, our regression model was created on the basis of assigning each selection criterion a value of 1.0 to generate a score for any potential patient. This assumes that each criterion is evenly weighted with respect to its association with significant carotid disease, which is not completely accurate, as indicated by the differences in relative risk presented in Table 1. However, our model's C-statistic of 0.82 (measurement of how well our model predicts observed outcomes, 1.0 indicating perfect predictability) attests to its accuracy.
Second, the advanced age of our subject population precluded evaluating age as a potential selection criteria for screening carotid sonography despite the known association of age and carotid disease [6, 9, 16]. We also failed to identify female sex as a selection criterion (risk factor) for significant carotid disease despite this known association [8, 9]. It is unclear why this relationship was not found; perhaps the number of subjects in our study was too small to allow this relationship to be identified, or possibly some selection criteria are universal whereas others depend on patient demographics.
In addition, our results apply strictly to subjects undergoing elective bypass. We excluded patients undergoing valve replacement (with or without bypass) or emergent coronary bypass because such patients are a completely different subset with respect to their risk factors for a postoperative cerebrovascular event.
Finally, we acknowledge that reviewing a patient's medical chart for selection criteria may be time-consuming and impractical. This is the reason we also obtained selection criteria directly from each patient and sided with the patient when any discrepancy between medical record and patient arose. Selection criteria should result from the most readily available source of information, which is usually the patient.
Despite these limitations, we believe we achieved our primary goal, to show the need and utility for some form of selection criteria when choosing patients to undergo carotid screening before elective CABGS.
In our study population, 24% of patients had significant ICA stenosis, which is in accord with other series [6, 8-10]. This also lends some credence to our results, indicating the utility of preselecting patients for a preoperative carotid sonography. Also, the relatively uniform prevalence of carotid disease in this targeted patient population allows an interesting observation. Consider the North American Symptomatic Carotid Endarterectomy Trial [1], in which 40% (1,124/2,814) of patients were found to have symptomatic heart disease. Hence, despite the logical thought process that coronary artery disease and carotid disease are linked, the link between confirmed carotid disease and suspected coronary disease seems to be stronger than that between confirmed coronary disease and suspected carotid disease. This observation, if not supportive of using selection criteria to identify patients who would benefit from preoperative carotid sonography, at least weakens the argument to perform carotid sonography in all individuals undergoing elective CABGS.
In conclusion, until the controversy regarding the need for preoperative carotid screening is resolved, we advocate and request the presence of at least one selection criterion when choosing patients who should undergo carotid sonography before elective CABGS. This approach markedly improves the probability of identifying a significant carotid stenosis on a screening examination without overlooking disease. Such a practice seems to logically follow from the existing extensive literature focused on identifying risk factors for carotid disease.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
