Embolic potential of left ventricular thrombi detected by two-dimensional echocardiography

PATHOPHYSIOLOGY AND NATURAL HISTORY THROMBOEMBOLIC DISEASE Embolic potential of left ventricular thrombi detected by two-dimensional echocardiography J. MARK HAUGLAND, M.D., RICHARD W. ASINGER, M.D., FRANK L. MIKELL, M.D., JOSEPH ELSPERGER, C.C.P.T., AND MORRISON HODGES, M.D. With the technical assistance of Lillian Michaud, R.D.M.S. ABSTRACT We sought to determine whether an association existed between the echocardiographic appearance of left ventricular thrombi and systemic embolization. We reviewed the clinical and echocardiographic characteristics of 60 patients who underwent diagnostic two-dimensional echocardiography for left ventricular thrombi. Sixteen of these 60 patients (27%) had evidence of systemic embolization. Multiple echocardiographic characteristics of left ventricular thrombi were analyzed, including mobility, shape, heterogeneity, echo density, layering, central echo lucency, presence within an aneurysm, and association with low-density swirling echoes. Incidence of embolization was significantly higher in patients with thrombi that were mobile or protruded into the left ventricular cavity (p <.002 and p <.05, respectively). Bayesian analysis indicated that the pretest likelihood for embolization was 27% and increased in the presence of mobility, central echo lucency, and protrusion to 60%, 50%, and 40%, respectively. A stepwise regression indicated that mobility was the first and protrusion the second most helpful echocardiographic characteristic in identifying patients with embolic phenomena. Clinical features were of less help in identifying the risk for embolization of patients with left ventricular thrombi. Nine of 31 patients (29%) with recent myocardial infarction (less than 3 weeks) had emboli in contrast to five of 26 patients (19%) with remote myocardial infarction (greater than 3 weeks) (p = NS). The three patients without infarction had congestive cardiomyopathy and two had emboli. Circulation 70. No. 4. 588-5989 1984. LEFT ventricular mural thrombus is a common postmortem finding in patients with acute myocardial infarction, '-5 remote myocardial infarction,1-' left ventricular aneurysm,46 and congestive cardiomyopathy.5 7 Clinical or pathologic evidence for embolization from left ventricular thrombus is much less common.'-' Since it appears that the incidence of embolic events can be reduced by therapeutic anticoagulation,j'9 it is important to identify patients with left ventricular thrombi so that therapeutic intervention can be considered. In the past, the diagnosis of left ventricular thrombus was based on contrast ventriculographic results or inferred only after clinical evidence for an embolic From the Cardiology Section, Department of Medicine, Hennepin County Medical Center and the University of Minnesota, Minneapolis. Address for correspondence: Richard W. Asinger, M.D., Non-Invasive Cardiology, Hennepin County Medical Center, 701 Park Ave., South, Minneapolis, MN 55415. Received May 9, 1983; revision accepted June 14, 1984. Presented in part at the 30th Annual Meeting of the American College of Cardiology, March 1981, San Francisco. 588 event. New techniques can reliably detect left ventricular thrombus in the absence of clinical evidence for embolization. Of these techniques, two-dimensional echocardiography is unique in that it not only is a sensitive and specific technique for detecting left ventricular thrombi'0 but also provides information regarding thrombus mobility, shape, and ultrasonic characteristics."-"5 Some of these characteristics, such as vigorous intracavitary motion and/or protrusion into the left ventricular cavity, appear ominous and have been reported to be associated with embolization. 16 We have reviewed our experience with left ventricular thrombi diagnosed by two-dimensional echocardiography to determine whether any echocardiographic characteristics were associated with systemic embolization. Materials and methods Patients. Over a 4 year period, during which we performed two-dimensional echocardiographic examinations in approximately 6000 patients, we identified 108 patients who demonstrated left ventricular echo-dense masses compatible with CIRCULATION

PATHOPHYSIOLOGY AND NATURAL HISTORY-THROMBOEMBOLIC DISEASE thrombi. The echocardiograms of these 108 patients were reviewed by two independent observers blinded to one another's interpretations for presence or absence of thrombi. Sixty patients had echocardiographic results meeting diagnostic criteria for left ventricular thrombus (class IV according to previous published criteria'5) by agreement of these two observers. The hospital records of these 60 patients were reviewed to determine the etiology of their heart disease, clinical indication for echocardiography, documentation of thrombus, anticoagulation status, and evidence for embolization before and after detection of thrombus. Echocardiographic technique and criteria for diagnosis of left ventricular thrombus. Two-dimensional echocardiographic examinations were performed with either a Varian V3000 electronic phased-array ultrasonoscope (Varian Associates, Salt Lake City) or an ATL Mark III mechanical sector scanner (Advanced Technology Laboratories, Bellevue, WA). A systematic examination was routinely performed. 17 Left ventricular thrombus was diagnosed if two-dimensional echocardiographic findings were characteristic of thrombus on a technically adequate examination.15 Characteristic features included the presence of an echo-dense mass in the left ventricle that was distinguishable from papillary muscles, muscle trabeculae, pseudotendons, 8 or tangenital views of the left ventricular wall. Clear delineation and distinction of the thrombus from the endocardium and myocardium was also required. Echocardiographic characteristics of thrombi. The echocardiograms of the 60 patients were reviewed a second time by the two echocardiographers for morphologic and acoustic characteristics of both the left ventricle and left ventricular thrombus. Classification of thrombi for these characteristics was made on the basis of agreement of the two echocardiographers. Interobserver agreement was defined for each characteristic as the percentage of patients in which the presence or absence of that characteristic was concurrent between the two observers. When disagreement occurred, the classification was made by a third observer. In addition, intraobserver agreement was determined for each characteristic after a second evaluation of all cases by one of the initial echocardiographers. Intraobserver agreement was defined as the percentage of concordance for each characteristic for both evaluations. Classifications were based on the first echocardiogram considered diagnostic for left ventricular thrombus. Intracavitarv motion. Three patterns of intracavitary motion were included in this classification. The first consisted of limited motion of a portion of thrombus with the same echo density as the remainder of the thrombus (figure 1). This pattern was limited to patients with protruding thrombi. The second pattern consisted of low~amplitude echoes with relatively free "wavy" motion in the left ventricular cavity but with a consistent base at a fixed thrombus (figure 2). This pattern in the experimental canine preparation of acute apical thrombus indicates the red cell and fibrin portion frequently referred to as the "red tail."''9 The third pattern consisted of a thrombus that was basically mural but demonstrated vigorous motion of its intracavitary margin during the cardiac cycle (figure 3). This particular characteristic was seen with left ventricular aneurysms that displayed marked dyskinetic motion and contained thrombi that frequently demonstrated central echo lucency. The relatively free intracavitary motion of low-amplitude dynamic swirling masses of echoes not based at a fixed thrombus was considered to represent the result of low flow,20 and hence these masses were not classified as thrombus. Thrombus shape. Thrombus shape was classified as either mural (figure 4) or protruding (figure 5). Flat thrombi parallel to the endocardial surface were classified as mural and those that projected into the left ventricular cavity as protruding. Vol. 70. No. 4, October 1984 B FIGURE 1. Modified apical four-chamber view (A) and accompanying diagram (B) demonstrating a protruding thrombus, which on real-time imaging displayed limited intracavitary motion. The arrow indicates the mobile portion of the thrombus. LV left ventricle; T = thrombus. = B FIGURE 2. Modified apical four-chamber view (A) and accompanying diagram (B) demonstrating low-density, low-amplitude, mobile intracavitary echoes consistently based at a fixed thrombus. Arrows indicate the mobile tail portion of the thrombus. Abbreviations as in figure 1. 59

HAUGLAND et al. B D FIGURE 3. Modified apical four-chamber view demonstrating vigorous motion of the intracavitary margin of a left ventricular thrombus. At end-systole (A and B) concavity of the thrombus margin with regard to the left ventricular chamber is demonstrated. At end-diastole (C and D) convexity of the thrombus margin with regard to the left ventricular chamber is evident. The difference in position of this thrombus margin between end-diastole and end-systole, indicated by the arrow, was obvious on real-time imaging and was associated with dyskinesis of the adjacent myocardium. Abbreviations as in figure 1. Echo-density characteristics. The echo density of each thrombus was classified as either of less echo density or of equal or greater echo density than adjacent myocardium. Each thrombus was also classified as either heterogeneous or homogeneous in echo density. Layering (figure 4) and evidence of central echo lucency (figure 6) were subclassifications of heterogeneous echo density. Left ventricular systolic perlcormance. A qualitative assessment of global systolic performance of the left ventricle was made for each patient. Global performance was classified as normal or decreased. Segmental performance was also noted for the area of myocardium in contact with the left ventricular thrombus. Segmental performance was classified as normal, hypokinetic, akinetic, or dyskinetic. Left ventricular aneurysm. Left ventricular aneurysm was defined qualitatively as an abnormality in the diastolic contour of the left ventricle (figure 4) with systolic dyskinesis. Clinical features. The 60 patients were classified according to the etiology of their heart disease. Fifty-seven had clinical and/or echocardiographic evidence of ischemic heart disease, while three had evidence of congestive cardiomyopathy. Those with ischemic heart disease were further classified according to the time from infarction to detection of left ventricular throm590 bus. Recent infarction was defined as that occurring within 3 weeks of detection of thrombus. Remote infarction was defined as that occurring more than 3 weeks before detection of thrombus. The anticoagulant status was also noted before and after detection of thrombus. Follow-up. Patients' charts were reviewed for clinical follow-up after detection of the thrombus. Follow-up end points included death, surgery to remove the thrombus, or longest recorded clinical follow-up. Statistical methods. The incidence of embolic phenomena for the various echocardiographic characteristics was tested with the chi-square test or Fisher's exact test.21 The sensitivity, specificity, and predictive value of these echocardiographic characteristics with regard to embolization were determined by the Bayesian theorem according to the nomenclature of Lusted.22 The echocardiographic characteristics were also analyzed with a stepwise regression23 to identify predictors for embolization. Differences were considered significant at p <.05. Results Clinical characteristics of patients (tables 1 and 2). Sixteen of the 60 patients had clinical events consistent CIRCULATION

PATHOPHYSIOLOGY AND NATURAL HISTORY-THROMBOEMBOLIC DISEASE B FIGURE 4. Apical four-chamber view (A) and accompanying diagram (B) demonstrating a left ventricular aneurysm with mural thrombus that also has evidence of layering. The apex demonstrated diastolic distortion and on real-time imaging marked dyskinesis. Note that an approximately 5 mm layer of the thrombus adjacent to the left ventricular cavity has increased echo density in comparison with the remainder of the thrombus. The arrow indicates the more echo dense outer layer of this thrombus. Abbreviations as in figure 1. with embolization (table 1). Pathologic documentation of embolization was available in six of the 16. Postmortem examination of four patients confirmed left ventricular thrombus and systemic emboli. Two patients had surgical removal of emboli from the aortic bifurcation and the left femoral artery. In the 10 patients without pathologic documentation of emboli, the diagnosis of embolization was made clinically. All had acute cerebrovascular events: five had multifocal abnormalities on neurologic examination and focal defects on computerized tomographic (CT) scans consistent with cerebral emboli. The remaining five patients had focal neurologic deficits. CT scan confirmed focal defects in three. One of the remaining two patients had a negative CT scan, but a focal abnormality was present on the electroencephalogram. The remaining patient did not have a CT scan but did have a focal abnormality on the electroencephalogram. Left ventricular thrombi were documented in 24 of the 60 patients: nine by postmortem examination, five at open heart surgery, and four by left ventriculography. The remaining six patients demonstrated develvol. 70, No. 4, October 1984 FIGURE 5. Modified apical long-axis view (A) and accompanying diagram (B) demonstrating a protruding left ventricular thrombus. Note that a portion of this thrombus is not parallel to the adjacent ventricular wall but projects into the left ventricular cavity. Abbreviations as in figure 1. FIGURE 6. Apical four-chamber view (A) and accompanying diagram (B) demonstrating central echo lucency of a left ventricular thrombus. At postmortem examination, this portion of thrombus was grossly liquid with scattered coarse fibrin strands. CL = central echo lucency; LV = left ventricle. 591

HAUGLAND et al. TABLE 1 Clinical and echocardiographic characteristics of patients with embolization Anticoag- Anticoagulation ulation Echo Central Clinical before after Follow-up Echo consis- echo Days from indi- LVT LVT period Patient IM Shape density tency Layering lucency Swirl Aneurysm MI to emb. cation detection detection (days) 1 No Pro. Homo No No No No No MI LVF None H-7d 28A 2 No Pro. Homo No No No No 20 EMB None H-2d 2A 3 Yes Pro. Homo No No No No 1 LVF None None 8A 4C No Pro ' Hetero Yes No No Yes 188 Serial None H-2d 2A 5D Yes Pro - Homo No Yes No Yes 17 EMB None H,C 368 6 No Pro < Homo No No No No 23 EMB None H,C 16 7 Yes Pro - Homo No No Yes No No MI EMB C H,C 71A 8 Yes Pro 2 Hetero Yes Yes No Yes 848 EMB None C 1630 9 No Flat < Hetero Yes Yes No Yes 13 EMB H,C H,C 45A 10 Yes Pro 2 Homo No No No No 4 LVF None H,C 809A 11 Yes Flat 2 Hetero Yes No No Yes 47 EMB H H,2d 2B 12 No Pro > Hetero No Yes No No 10 EMB None H,C 39 13 Yes Pro 2 Homo No No No Yes Recent MI, EMB None H,C 410 age indeter. 14 Yes Flat 2 Hetero Yes No No Yes 45 EMB None H,C 29 15 Yes Flat < Hetero No No No No 8 EMB None None 19A 16 No Pro < Hetero No No No No 4 EMB H-4d H,C 36B IM = intracavitary motion; Pro protrusion into left ventricular cavity; - = echo density equal to or greater than that of myocardium; < = echo density less than that of myocardium; Homo = homogeneous echo density; Hetero = heterogeneous echo density; MI = myocardial infarction; LVF = left ventricular function; EMB = embolic source; indeter. = precise time indeterminant; serial = echocardiography performed as part of an ongoing prospective study of left ventricular thrombus in a consecutive group of patients with acute myocardial infarction; LVT = left ventricular thrombus; H - heparin for indicated number of days (d); H,C = heparin followed by coumadin for at least 1 mo or until follow-up end point. AFollow-up to death. BFollow-up to surgery. CPatient had myocardial infarction complicated by thrombus 188 days before embolization treated initially with anticoagulants, but these were discontinued 3 mo before readmission for heart failure. At readmission, thrombus was again detected and heparin was initiated. The following day embolization to bowel occurred. Surgical intervention was refused and the patient died the next day. DPatient had perioperative infarction and embolization to lower extremities 17 days after surgery. Thrombus detected at that time and anticoagulants were initiated. Recurrent embolization occurred 7 days later. opment or resolution of thrombi on serial echocardiograms. Of the 57 patients with ischemic heart disease, 31 had left ventricular thrombus diagnosed within 3 weeks of their infarction; embolization occurred in nine (29%). Only five (19%) of the 26 patients with remote infarction had an embolic event (p = NS). Two of the three patients with congestive cardiomyopathy had embolic events. Echocardiographic characteristics Thrombus motion. Fifteen patients had left ventricular thrombi demonstrating intracavitary motion; nine (60%) had embolic events. Forty-five patients had no evidence of intracavitary motion; seven (16%) had embolic events. The incidence of emboli was significantly higher for thrombi that displayed intracavitary motion (p <.002) (table 3). Interobserver agreement for this characteristic was 82%; intraobserver agreement was 92%. 592 Intracavitary motion was the single most important characteristic to increase the posttest likelihood of embolization (table 4). The pretest likelihood of embolization given the presence of a left ventricular thrombus was 27% (16/60). The presence of intracavitary motion increased the posttest likelihood of embolization to 60%. In addition, intracavitary motion was chosen as the most significant predictive parameter for embolization in the stepwise regression (R2 =.19, p <.005) (table 5). Low-amplitude swirling intracavitary echoes consistent with low flow were present in nine patients; only one had evidence of embolization (11%). Interobserver agreement for this characteristic was 95%, and intraobserver agreement was 90%. Fifteen of 51 patients without this characteristic had embolization (29%; p = NS). The absence of swirling intracavitary echoes was sensitive in identifying patients with embolization but lacked specificity (table 4) and did not CIRCULATION

PATHOPHYSIOLOGY AND NATURAL HISTORY-THROMBOEMBOLIC DISEASE significantly increase the posttest likelihood for embolization. With the stepwise regression analysis the absence of swirling echoes, in addition to the parameters of intracavitary motion, protrusion, and central echo lucency, increased the R2 value from.24 to.27 (p = NS) (table 5). Thrombus shape. Twenty-nine patients had protruding left ventricular thrombi; 12 (41%) had embolic events. In contrast, 31 patients had flat mural thrombi; four (13%) had embolic events. The incidence of embolic phenomena was significantly higher with protruding than with flat thrombi (p <.05). Interobserver agreement for this characteristic was 88%, and intraobserver agreement was 93%. Protrusion of the thrombus into the left ventricular cavity increased the likelihood for embolization to a posttest value of 40% (table 4). By stepwise regression, protrusion was chosen as the second most valuable echocardiographic characteristic for predicting embolization. The addition of protrusion to intracavitary motion increased the R2 value from.19 to.22 (p = NS) (table 5). Echo-density characteristics DENSITY. The echo density of the thrombus was greater than or equal to that of the surrounding myocardium in 41 patients; 12 had embolic events (29%). Of the 19 patients with echo density less than that of the surrounding myocardium, four had evidence for embolization (21 %; p = NS). Interobserver agreement for this characteristic was 83%, and intraobserver agreement was 85%. The echo density of a thrombus did not significantly change the likelihood for embolization (table 4). In addition, this parameter was not chosen in the stepwise regression as being helpful in determining the risk of embolization. HETEROGENEOUS ECHO DENSITY. There were 32 patients with heterogeneous echo density; eight had embolic events (25%). Twenty-eight patients had homogeneous-appearing thrombi; eight had emboli (29%; p = NS). Interobserver agreement for this characteristic was 58%, and intraobserver agreement was 82%. Homogeneity of the thrombus did not significantly change the likelihood for embolization (table 4) and was not chosen in the stepwise regression as being helpful in determining the risk of embolization. Heterogeneity in echo density was secondary to a layered pattern in 19 patients; five had embolic events (26%). Of the 41 patients without a layered pattern, 11 had embolic events (27%; p = NS). Interobserver agreement for this characteristic was 87%, and in- Vol. 70, No. 4, October 1984 traobserver agreement was 73%. Layering of the thrombus did not significantly change the likelihood for embolization (table 4) and was not chosen in the stepwise regression as being helpful in determining the risk of embolization. Heterogeneity in echo density was secondary to central echo lucency in eight patients, four of whom had embolic events (50%). Of the 52 patients without central liquefaction, 12 had embolic events (23%; p = NS). Interobserver agreement for this characteristic was 97%, and intraobserver agreement was 90%. The presence of central echo lucency increased the likelihood of embolization from 26% to 50% (table 4). When this characteristic was added to intracavitary motion and protrusion, the R2 value increased from.22 to.24 (p = NS) (table 5). Left ventricular systolic performance. All 60 patients with left ventricular thrombi had decreased systolic performance of the left ventricle. Fifty-five of the 57 patients with ischemic heart disease had apical thrombi and all had severe apical wall motion abnormality (i.e., either akinesis or dyskinesis). Two patients with ischemic heart disease had thrombi located along the inferior wall. The wall motion adjacent to the thrombus was either akinetic or dyskinetic in both cases. All patients with congestive cardiomyopathy had decreased global systolic performance of the left ventricle. Thrombus was located near the apex in each case and was associated with apical hypokinesis. Left ventricular aneurysm. Twenty-five patients had echocardiographic evidence of left ventricular aneurysm, with thrombi confined to the aneurysm in each case; seven (28%) had emboli. Of the 35 patients without left ventricular aneurysm, nine (26%) had emboli (p = NS). Interobserver agreement for this characteristic was 78%, and intraobserver agreement was 77%. The presence of a left ventricular aneurysm did not increase the likelihood for embolization (table 4). Likewise, this parameter was not chosen in the stepwise regression as being helpful in determining the risk of embolization. Follow-up. Follow-up of patients after detection of left ventricular thrombus was 296 ± 413 days (range 1 to 1630). Of the patients who did not die or undergo surgical removal within 2 weeks, follow-up exceeded 2 weeks in all but three. Although only five patients were on therapeutic anticoagulants before echocardiographic detection of thrombi, 44 received anticoagulants after detection of thrombi. The duration of anticoagulant therapy after detection of thrombus varied; those indicated as receiving anticoagulants in tables 1 and 2 were on anticoagulant therapy at least 1 month or 593

HAUGLAND et al. TABLE 2 Clinical and echocardiographic characteristics of patients without embolization Days Anticoagulation status Echo Central from Prior to After Follow-up Echo consis- Layer- echo MI to Clinical LVT LVT period Patient IM Shape density tency ing lucency Swirl Aneurysm echo indication detection detection (days) 1 Yes Pro > Hetero No No No No 21 LVF None H,C 1 1 A 2 No Pro. Homo No No Yes Yes 38 LVF None C 608 3 No Flat < Hetero Yes No No Yes 98 Serial None H,C 854 4 No Flat < Homo No No No No 89 Serial None H,C 836A 5 No Flat - Hetero Yes No No No 335 Serial None None 430 6 No Flat - Hetero No No No No 7 Serial None H,C 1288 7 Yes Pro < Hetero No No No Yes 8 Serial None H,C 112 8 Yes Pro 2 Homo No No No No No MI LVF None H,C 2l4A 9 Yes Pro - Homo No No Yes Yes 8 LVF None H,C 418 10 No Flat 2 Homo No No No No 7 Serial None None 5A 11 No Flat - Hetero No No No Yes 83 Serial None None 420 12 No Flat 2 Hetero Yes No No No 14 LVF None None 528 13 No Flat < Homo No No No No Remote LVF None H,C 92A indeter. 14 No Flat < Hetero No Yes Yes Yes 300 Effusion None None 7A 15 No Flat - Hetero Yes No No Yes 80 Serial None Hc 164 16 No Flat - Hetero Yes No No Yes 30 LVF None H,C 46A 17 No Pro < Hetero No No No No Remote LVF None H,C 15 indeter. 18 No Flat < Homo No No Yes Yes 30 Serial None H,C 432 19 No Flat < Homo No No No No 2 Serial None C 1430 20 No Pro < Homo No No No No 5 Serial None H,C 83A LVT = left ventricular thrombus; IM = intracavitary motion; Pro protrusion into LV cavity; -= echo density equal to or greater than that of myocardium; < = echo density less than that of myocardium; Homo = homogeneous echo density; Hetero = heterogeneous echo density; MI - myocardial infarction; LVF = left ventricular function; EMB = embolic source; AOVD = aortic valve disease; indeter. = precise time of infarct indeterminant; serial = echocardiography performed as part of an ongoing prospective study of left ventricular thrombus in a consecutive group of patients with acute myocardial infarction; H = heparin; C = coumadin. AFollow-up to death. BFollow-up to surgery. CPatient refused coumadin. DPatient had cerebrovascular accident 1 day after MI. Echocardiography 90 days after MI demonstrated left ventricular thrombus; followed 52 additional days without emboli. EPatient had cerebrovascular accident 94 days before echocardiography demonstrating left ventricular thrombus; followed 167 additional days without emboli. FProtrusion in this patient limited to protrusion into surgically confirmed pseudoaneurysm. until a follow-up end point (death, surgery, or longest four of these five had embolization within 1 day of follow-up). detection, the anticoagulant status after detection of Five of the 16 patients with embolization had emboli thrombus was not included as a parameter in any statisafter echocardiographic detection of thrombi (patients tical analysis. 3, 4, 5, 10, and 11; table 1). One patient had an Of the five who had embolic events after detection embolic event within hours of echocardiographic ex- of thrombus, four had thrombi that demonstrated intraamination. Three had embolic events the day after cavitary motion and four had protruding thrombi. The echocardiographic examination, and the fifth had em- one flat thrombus that embolized displayed intracavibolization 7 days after initial detection of thrombus. tary motion. Only one of these patients had a thrombus Three of these five had had emboli before echocardio- with central echo lucency and none had swirling intragraphic detection of thrombus. Two of these three with cavitary echoes. recurrent emboli were on intravenous heparin at the time of the second clinically apparent embolic event. Discussion Because the number of patients who had embolic The results of this retrospective analysis of embolic events after detection of thrombus was low and since phenomena and the echocardiographic characteristics 594 CIRCULATION

PATHOPHYSIOLOGY AND NATURAL HISTORY-THROMBOEMBOLIC DISEASE TABLE 2 (Continued) Days Anticoagulation status Echo Central from Prior to After Follow-up Echo consis- Layer- echo MI to Clinical LVT LVT period Patient IM Shape density tency ing lucency Swirl Aneurysm echo indication detection detection (days) 21 No Flat 2 Hetero Yes No No Yes 12 Serial None None 47A 22 No Flat < Hetero Yes No No Yes 60 Pseudo- None None 2B aneurysm 23 No Flat 2 Hetero Yes No No No 7 Serial None H,C 177 24 No Pro > Hetero No No Yes No 11 Serial None H 6A 25 No Pro - Hetero No No No No 4 Serial None H,C 95 26 Yes Pro - Homo No No No No 1 LVF None H 5A 27 No Flat < Homo No No No No 3 LVF None None 28 No Flat < Homo No No Yes No 90 EMBD C C 52 29 No Pro < Homo No No No Yes 4 Serial None H,C 905 30 No Pro 2 Hetero Yes No Yes Yes 1 Serial None H,C 601 31 No Flat 2 Hetero No No No No 76 Serial None C 14A 32 No Flat 2 Hetero Yes Yes No Yes 126 LVF None None 10 33 No Pro - Homo No No Yes No Remote LVF None H,C 494 indeter. 34 Yes Pro 2 Hetero Yes Yes No No 5 Pseudo- None H,C 686 aneurysm 35 No Flat 2 Homo Yes No No Yes 10 Serial None H,C 103 36 No ProF 2 Hetero No No No Yes 90 LVF None None 2B 37 No Flat 2 Homo No No No No 4 Serial None H,C 8 38 No Flat < Homo No No No Yes 60 Serial None C 1367 39 No Pro - Homo No No No No 5 Effusion None None 403 40 No Flat < Homo No No No No 1 Serial None H,C 1034 41 No Pro - Hetero No No No No Remote AOVD None None 27 indeter. 42 No Flat. Hetero Yes No No No 7 Serial None H,C 8 43 No Pro 2 Homo No No No No 94 EMBE None None 167 44 No Flat 2 Hetero Yes Yes No Yes 108 LVF None None 60B of left ventricular thrombi indicate that mobile thrombi are more likely to be associated with embolization than are immobile thrombi. Similarly, thrombi that protrude into the left ventricular cavity are more likely to be associated with embolization than are flat thrombi. However, thrombi with mobility and protrusion do not represent a significantly higher risk group for embolization than do mobile thrombi alone. These observations are based on the results of a retrospective study in which several of the patients had suffered an embolic event before echocardiographic evaluation and were being examined to establish an intracardiac source. Hence our results do not necessarily establish the prospective risk for embolization given the presence of left ventricular thrombi with the characteristics described. Two-dimensional echocardiography provides a sensitive and specific noninvasive method to detect left ventricular thrombi.'>'5 In addition, this technique provides unique data regarding mobility, shape, and acoustic characteristics in vivo that cannot be obtained with other currently used techniques. Our findings indicate that some of these unique characteristics show a significant association with embolization. Mobility of a portion of the thrombus within the left ventricle was the characteristic most significantly associated with embolization. DeMaria et al.'2 first described the association of this characteristic with embolization. Subsequent reports'6 24' 25 and our results indicate an association between mobility of thrombi and embolization. Distinction of the various types of mobile intracavitary echoes associated with thrombi is important. Mobility of a portion of a protruding thrombus that has the same echo density as the remainder of the thrombus is easily detected. These mobile portions usually have a broad base on the fixed thrombus. A second type of intracavitary motion consists of distinct low-amplitude echoes, again based at a fixed thrombus. The mobile portion demonstrates a wavy motion within the left ventricular cavity. This type of intra- Vol. 70, No. 4, October 1984 595

HAUGLAND et al. TABLE 3 Incidence of embolization in relation to echocardiographic characteristics of left ventricular thrombi Incidence of Incidence of embolization embolization Echocardiographic in presence of in absence of characteristic characteristic characteristic p value Intracavitary motion 9/15 7/45 (60%) (16%).002 Protrusion 12/29 4/31 (41%) (13%).05 Central echo lucency 4/8 12/52 (50%) (23%) NS Equal or increased 12/41 4/19 echo density (29%) (21%) NS Aneurysm 7/25 9/35 (28%) (26%) NS Homogeneous 8/28 8/32 (24%) (25%) NS Layering 5/19 11/41 (26%) (27%) NS Swirl 1/9 15151 (11%) (29%) NS cavitary motion is unusual in the clinical setting; however, it is frequently noted in an experimental canine preparation of left ventricular apical thrombus'9 and corresponds histologically with the "'red tail" portion of the thrombus. The third pattern of intracavitary motion included in this study was vigorous motion of the intracavitary margin of the thrombus. This pattern was noted primarily in patients with markedly dyskinetic areas and was frequently associated with central echo lucency of the thrombus. Disruption of this mobile intracavitary margin with release of its thrombotic contents into the systemic arterial circulation could account for its association with embolization. Another type of mobile intracavitary pattern noted TABLE 4 Bayesian analysis: sensitivity, specificity, and posttest likelihood of embolism for various echocardiographic characteristics of left ventricular thrombi Posttest likelihood Echocardiographic Sensitivity Specificity of embolism characteristic (%) (%) (%) Intracavitary motion 56 86 60 Central echo lucency 25 91 50 Protrusion 75 59 40 Absence of swirl 94 18 29 Homogeneous 50 55 29 Equal or increased echo density 75 33 29 Aneurysm 44 59 28 Absence of layering 69 32 27 596 with or without thrombi consists of low-amplitude, low-density echoes within the left ventricular cavity. This type of echo pattern differs from the "red tail" pattern in that it is characterized by less echo density, moves freely within the left ventricular cavity, and has no apparent attachment to a discrete thrombus. This echo pattern is noted in areas of low flow, such as dilated left ventricles with very poor systolic performance or large left ventricular aneurysms. It can be produced in a short-term experimental canine preparation of apical infarction and does not indicate formation of thrombus.20 Nine patients in this report demonstrated such intracavitary echoes but only one had an embolic event. Hence, this echocardiographic pattern was unusual for patients with embolic phenomena. Perhaps thrombi located in areas of low flow are not as likely to embolize as those located in areas where blood flow is more rapid. Protruding thrombi were also associated with embolization in this series. Cabin and Roberts26 postulated that thrombi that protrude into the left ventricular cavity have relatively more surface contact with mobile intracavitary blood and hence are more likely to embolize. They used this concept to explain their observation of a lower incidence of embolization for patients with left ventricular thrombus contained within an aneurysm than that for patients with congestive cardiomyopathy and protruding thrombus. Our findings demonstrate an association between protruding thrombi and embolization but do not support a protective role of left ventricular aneurysm against embolization. However, there may be differences in left ventricular aneurysm defined by postmortem examination and by echocardiographic findings that could account for the differences in our observations. We defined left ventricular aneurysm qualitatively as an area of diastolic distortion and systolic dyskinesis. Cabin and Roberts26 defined left ventricular aneurysm at postmortem examination as an area of thinned myocardium with obvious distortion on gross external examination. TABLE 5 Stepwise regression of echocardiographic characteristics for embolization Echocardiographic characteristic R2 p value Intracavitary motion.19 <.001 Protrusion.22 <.25A Central echo lucency.24 <. 15A Absence of swirl.27 <.15A APartial p value indicating probability that additional parameter adds to previously selected parameters in predicting embolization. CIRCULATION

PATHOPHYSIOLOGY AND NATURAL HISTORY-THROMBOEMBOLIC DISEASE We27 and others28 have observed the development of diastolic distortion and dyskinesis in areas of infarction during the first days and weeks after acute myocardial infarction. It has been our experience that patients with infarction who develop these echocardiographic findings and die do not necessarily have postmortem findings typical of aneurysm. Hence, differences in left ventricular aneurysm as defined echocardiographically and by postmortem examination may explain why we did not demonstrate a protective role of left ventricular aneurysms for embolization. We did, however, note that the low-density swirling intracavitary echoes, which we have associated with low flow states,20 were rather frequent in patients with large left ventricular aneurysms that contained thrombus. Perhaps the flow characteristics within an aneurysm are important factors with regard to embolization in addition to the shape of the thrombus. Of the other echocardiographic characteristics analyzed, only central echo lucency aided in distinguishing patients with evidence of embolization. This echocardiographic pattern was frequently associated with vigorous motion of the intracavitary margin of a flat thrombus. In our one patient with central echo lucency in whom pathologic data were obtained at postmortem examination, the central echo-lucent portion of the thrombus was liquid rather than solid. Since thrombi generally form in an area of severe wall motion abnormality, a change in the physical property of a portion of the thrombus from solid to liquid may allow dyskinetic motion to become more apparent along the intracavitary border. It is possible that this increased mobility of the intracavitary thrombus border could lead to disruption and embolization. Such a mechanism could account for the association of these characteristics with embolization. In our patients, thrombi detected within 3 weeks of infarction were more likely to be associated with embolic events than those detected later than 3 weeks after infarction, an observation consistent with previous findings.2 It is important to note, however, that many of our patients underwent echocardiographic examination as part of a separate, ongoing serial study on the development of left ventricular thrombus after acute infarction.27 Left ventricular thrombus was detected early in the course of infarction in these patients and, in general, they were treated with therapeutic anticoagulation immediately upon diagnosis of thrombus. Their subsequent course is not necessarily representative of the natural history of left ventricular thrombus after infarction, particularly with regard to embolic risk. As two-dimensional echocardiography becomes a more widely available and utilized technique, a frequent clinical problem will be the discovery of unsuspected left ventricular thrombi. In this situation the clinician is faced with the problem of assessing the embolic potential of such thrombi. Our findings may provide some insight into the embolic potential of these thrombi. In addition to the clinical setting, certain echocardiographic characteristics of thrombi may be helpful in determining the risk for embolization. Mobility, shape, and central echo lucency appear to be the most significant echocardiographic features associated with embolization, and of the current invasive or noninvasive procedures available, two-dimensional echocardiography is the easiest and most reliable method to detect these characteristics. These features should be considered in addition to the clinical findings when assessing the risk of embolization. We gratefully acknowledge the secretarial services provided by Belinda Anderson. Darryl A. Erlien and Dr. George Logan provided the statistical analysis. References 1. Bean WB: Infarction of the heart. II1. Clinical course and morphologic findings. Ann Intern Med 12: 71, 1938 2. Nay RM, Barnes AR: Incidence of embolic or thrombotic processes during the immediate convalescence from acute myocardial infarction. Am Heart J 30: 65, 1945 3. Jordan RA, Miller RD, Edwards JE, Parker RL: Thrombo-embolism in acute and in healed myocardial infarction. I. Intracardiac mural thrombosis. Circulation 6: 1, 1952 4. Yater WM, Welsh PP, Stapleton JF, Clark ML: Comparison of clinical and pathologic aspects of coronary artery disease in men of various age groups: a study of 950 autopsied cases from the Armed Forces Institute of Pathology. Ann Intern Med 35: 352, 1947 5. Garvin CF: Mural thrombi in the heart. Am Heart J 21: 713, 1941 6. Phares WS, Edwards JE, Burchell HM: Cardiac aneurysms: clinicopathologic studies. Staff Meeting Mayo Clin 28: 264, 1953 7. Roberts WC, Ferrans VJ: Pathologic anatomy of the cardiomyopathies: idiopathic dilated and hypertrophic types, infiltrative types, and endomyocardial disease with and without eosinophila. Human Pathol 6: 287, 1974 8. Ebert RV: Anticoagulants in acute myocardial infarction: results of a cooperative clinical trial. JAMA 225: 724, 1973 9. Report of the working party on anticoagulant therapy in coronary thrombosis to the medical research council: Assessment of shortterm anticoagulant administration after cardiac infarction. Br Med J 1: 335, 1969 10. Stratton JR, Lighty GW Jr, Pearlman AL, Ritchie JL: Detection of left ventricular thrombus by two-dimensional echocardiography: sensitivity, specificity, and causes of uncertainty. Circulation 66: 150, 1982 11. Ports JA, Cogan J, Schiller NB, Rapaport E: Echocardiography of left ventricular masses. Circulation 58: 528, 1978 12. DeMaria AN, Bommer W, Neumann A, Grehl T, Weinart L, De Nardo S, Amsterdam EA, Mason DT: Left ventricular thrombi identified by cross-sectional echocardiography. Ann Intern Med 90: 14, 1979 13. Meltzer RS, Gutarer D, Rakowski H, Popp RL, Martin RD: Diagnosis of left ventricular thrombi by two-dimensional echocardiography. Br Heart J 42: 261, 1979 14. Reeder GS, Tajik AJ, Seward JB: Left ventricular mural thrombus: two-dimensional echocardiographic diagnosis. Mayo Clin Proc 56: 82, 1981 Vol. 70, No. 4, October 1984 597

HAUGLAND et al. 15. Asinger RW, Mikell FL, Sharma B, Hodges M: Observations on detecting left ventricular thrombus with two-dimensional echocardiography: emphasis or avoidance offalse positive diagnosis. Am J Cardiol 47: 145, 1981 16. Visser CA, Kan G, Meltzer RS, Roelandt J, Durrer D: Twodimensional echocardiographic features of left ventricular thrombi causing peripheral embolization. Circulation 68 (suppl III): III- 1 11, 1983 (abst) 17. Rourke T, Asinger RW, Mikell FL, Francis GS, Hodges M: Sector echocardiographic technique: a systematic approach. Med Ultrasound 3: 51, 1979 18. Nishimura T, Kondo M, Umadome H, Schimono Y: Echocardiographic features of false tendons in the left ventricle. Am J Cardiol 48: 177, 1982 19. Mikell FL, Asinger RW, Elsperger KJ, Anderson WR, Hodges M: Tissue acoustic properties of fresh left ventricular thrombi and visualization by two-dimensional echocardiography: experimental observations. Am J Cardiol 49: 1157, 1982 20. Mikell FL, Asinger RW, Elsperger KJ, Anderson WR, Hodges M: Regional stasis of blood in the dysfunctional left ventricle: echocardiographic detection and differentiation from early thrombosis. Circulation 66: 755, 1982 21. Siegel S: Non-parametric statistics for the behavioral sciences. New York, 1956, McGraw-Hill Book Co. 22. Lusted LB: Introduction to medical decision making. Springfield, IL, 1968, Charles C Thomas Publisher 23. Jennrich RI: Stepwise regression. In Enslein K, Ralston A, Wilf HS editors: Statistical methods for digital computers. New York, 1977, John Wiley and Sons, Inc., vol Ill, pp 58-75 24. Lenen RF, Vidne B, Sclarousky S, Aomon J: Two-dimensional real-time echocardiographic detection of a left ventricular aneurysm with mobile pedunculated thrombi. Chest 77: 704, 1980 25. Nagvi N, Popp RL, Coltart DJ: Diagnosis of left ventricular thrombosis by two-dimensional echocardiography. Eur J Cardiol 11: 235, 1980 26. Cabin HS, Roberts WC: Left ventricular aneurysm, intra-aneurysmal thrombus, and systemic embolus in coronary heart disease. Chest 77: 586, 1980 27. Asinger RW, Mikell FL, Esperger J, Hodges M: Incidence of left ventricular thrombosis after acute transmural myocardial infarction. N Engl J Med 305: 297, 1981 28. Eaton LW, Weiss JL, Bulkley BH, Garrison JB, Weisfeldt ML: Regional cardial dilitation after acute myocardial infarction: recognition by two-dimensional echocardiography. N Engl J Med 300: 57, 1979 598 CIRCULATION