We assessed the histologic features of 526 carotid plaques from consecutive patients undergoing endarterectomy for symptomatic carotid stenosis and found a high prevalence of coronary-type plaque instability, with strong correlations between macrophage infiltration and both cap rupture and time since stroke. Temporal trends were much weaker after a transient ischemic attack than after a stroke, with a tendency for plaque features to persist for a longer period, suggesting heterogeneity in the underlying pathological mechanisms.
Thrombus formation in moderately stenosed coronary arteries at the location of an eroded or ruptured fibrous cap is the cause of most acute coronary syndromes.1 A substantial proportion of ischemic strokes are also caused by atherosclerotic plaque but the mechanisms are less well understood and “stable” plaque causing cerebral hypoperfusion is sometimes responsible.
Previous studies have compared histological features in symptomatic and asymptomatic plaques, but most of these studies have been small (a recent systematic review2 found a median of 43 plaques per study) and results have been inconsistent.2 Larger studies of plaque macroscopic appearances3,5 or plaque appearances on imaging are difficult to interpret because it is not clear which pathological features are represented by these indirect assessments.
Subjects and methods
We studied the carotid plaques of consecutive patients with symptomatic carotid stenosis who underwent carotid endarterectomy (CEA). Baseline characteristics and the nature and timing of ipsilateral cerebral or ocular ischemic symptoms were recorded by a study neurologist. Stroke was defined as ischemic symptoms lasting longer than 24 hours.
We fixed the excised plaques in formalin and embedded them in paraffin wax. We then took 5 µm thick sections from the bifurcation and from 3 mm on either side and stained them with hematoxylin and eosin, elastin van Gieson, and CD68 and CD3 antibody to identify macrophages and lymphocytes, respectively. We used semiquantitative scales6,7 to measure surface thrombus, intraplaque hemorrhage, plaque and cap infiltration, vascularity, foam cells, lipid core, and cap rupture. Cap thickness was measured with a calibrated graticule in the microscopic eyepiece. A minimum cap thickness of < 200 µm was classified as a “thin cap,” as this was the median value recorded in our study. We also categorized each plaque according to the American Heart Association (AHA) scheme8 and our own composite assessment of overall plaque instability (stable, predominantly stable, unstable with intact cap, and unstable with ruptured cap). The aim of this classification was to include all potential markers of instability and was based on widely accepted definitions of unstable plaque in the coronary circulation.1
The prevalence of histologic features was related to the timing and nature of the most recent ischemic event using the chi-square test and logistic regression models in which the effect of time since symptoms was represented as a cubic spline with a single knot.9
Five-hundred-twenty-six plaques from 498 consecutive patients (72.1% male, mean ± SD age 66.6 ± 8.7 years) were studied. The most recent ischemic event was a stroke in 159 patients (hemispheric, 135; ocular, 24), and a transient ischemic attack (TIA) in 367 patients (hemispheric, 201; ocular, 166). Patients with stroke were less likely to have had multiple symptomatic ischemic episodes before surgery and to have experienced a greater delay between the most recent ischemic event and CEA than the patients with TIA (median, 126 days vs 67 days; < .001). Other baseline characteristics were similar in the symptom subgroups (
Most plaques had marked inflammation (66.8%), intraplaque hemorrhage (64.6%), and a ruptured cap (58.1%); 50.4% of plaques were classified as AHA Grade 6, and 64.1% were unstable. Histologic features associated with a ruptured fibrous cap are shown in
There was a significant inverse relationship between some histologic features and the time since stroke, including “unstable” plaque ( = .001), cap rupture ( = .02), and plaque macrophages ( = .007;
). Much weaker associations were seen between these features and time since TIA. Plaques removed within 7 days after a TIA tended to be more unstable on histologic examination than plaques removed 8 to 30 days after a TIA (eg, unstable plaque, 74.5% vs 60.5%, respectively, = .12; marked plaque macrophages, 75% vs 60%, respectively, = .09), but patients who underwent surgery within 7 days after a TIA were more likely to have multiple ischemic events, with the most recent event taking place after the date for surgery was set. For the 67 patients with only 1 TIA prior to CEA, no temporal trends in plaque features were noted.
The results of our study showed that among the 526 symptomatic carotid plaques assessed, there was a high incidence of marked inflammation, large lipid core, and ruptured cap, indicating that the mechanisms of carotid plaque instability are similar to those in culprit coronary plaques.1,8 The dense macrophage infiltration related to time since stroke and cap rupture suggests a causative association between carotid plaque instability and inflammation. Plaque vascularity and lymphocyte infiltration, however, were not consistently related to time since stroke or cap rupture.
Our findings add significantly to those of previous histological studies of carotid plaque, most of which simply compared symptomatic and asymptomatic plaques.2 First, whilst previous studies have shown that macrophage infiltration is greater in symptomatic plaques compared to asymptomatic plaques,10, 11 we have shown that macrophage infiltration is strongly associated with both cap rupture and time since stroke. Second, in contrast to the time trends observed in plaques from patients with stroke, unstable plaque features tended to persist for a longer period after a TIA. Thus the form of plaque instability that leads most frequently to TIA may be more chronic than that leading most frequently to stroke in patients with severe carotid stenosis. Interestingly, plaques from patients with TIA tended to have a greater prevalence of surface thrombus in the absence of cap rupture (OR = 1.62 [0.77-3.51], = .17) which may reflect a higher prevalence of plaque erosion in patients with TIA.
If the pathological mechanisms underlying stroke and TIA in patients with carotid stenosis truly are different, then one might expect the clinical course to differ between patients with TIA versus stroke. We found that patients with TIA often had multiple ischemic episodes in the months prior to CEA, whereas patients with stroke tended to have an isolated episode. Thus there may be a spectrum of plaque pathology leading to distinct clinical patterns. At one end of the spectrum, acute plaque destabilization results in an isolated severe ischemic episode followed by “healing” to a relatively quiescent state over several months, perhaps mediated by a shift in the balance of matrix-degrading enzymes and their inhibitors within the plaque.12 At the other end of the spectrum, a more chronic process results in repeated, less severe ischemic symptoms over a longer period of time, perhaps suggesting that emboli are smaller or different in composition.
Other possible reasons for our results exist. Because patients with TIA may be more likely to have asymptomatic plaque embolization than those with stroke13 or because they may have less precise recall, the information on timing of TIAs may be less accurate than that for strokes. In addition, some acute consequences of the stroke, which may be less marked or absent after a TIA, may be responsible for plaque inflammation. This reverse-causation seems unlikely, however.
Our study had several potential limitations. It is possible that we missed some histologic features that may have been present between the sections we took at 3-mm intervals along the length of each plaque. In a pilot study we showed that there was good agreement between adjacent 3 mm sections and that even a single bifurcation section was reasonably representative of the plaque as a whole.6 In addition, because CEA was often delayed following a stroke, the power to detect transient changes in pathological features in patients with stroke was limited. A postmortem study of acutely occluded carotid arteries in patients with fatal stroke, however, showed that plaque histology was comparable with that seen in our patients, who underwent surgery within 60 days.14 Furthermore, a study of intracoronary thrombectomy specimens removed within 6 hours of ST-elevation myocardial infarction found that features of unstable plaque had often been present for several weeks.15
Our study showed a high prevalence of coronary-type plaque instability in the carotid plaques of patients who underwent CEA for symptomatic carotid stenosis, with strong correlations between macrophage infiltration and both cap rupture and time since stroke. Temporal trends were much weaker after a TIA than after a stroke, with a tendency for plaque features to persist for a longer period, suggesting heterogeneity in the underlying pathological mechanisms.