Although cardiac myxoma is occasionally asymptomatic, 3 major categories of presentation are common: cardiac symptoms due to atrioventricular valve obstruction, neurologic manifestations caused by systemic embolization (peripheral or cerebral), and constitutional symptoms.^1,2 Cardiac symptoms, which occur in 60% to 90% of all patients, most often provoke diagnosis.^1–3 When neurologic complications—such as stroke or transient ischemic attack (TIA) due to cerebral embolism—present as the initial manifestation, this subset of patients is usually first treated by physicians other than cardiologists or cardiac surgeons. This can delay the diagnosis of myxoma and increase the risk of sequelae. In fact, central embolisms occur more often than expected in the setting of cardiac myxoma—in 10% to 30% of all cases.^1,2 Left atrial myxoma has been well documented in the medical literature, but few of those studies have focused on central nervous system manifestations attributed to myxoma. Are there any differences in clinical presentation between patients with embolic sequelae and those without? Does neurologic damage have an influence on the diagnostic process, the timing of surgery, and the long-term outcome? The above questions should all have common-sense answers that already are widely known. However, few papers from large clinical series (most of them case reports) explain this “old idea” with data. We now report what is, to the best of our knowledge, the largest clinical case series of cerebral manifestations of left atrial myxoma to come from a single institution.

Patients and Methods

In performing a search of hospital records at Tongji Hospital (a tertiary-care teaching hospital in central China), we identified 92 patients who had been diagnosed with cardiac myxoma from January 2000 through December 2012. Of these, 5 were excluded because their diagnoses of cardiac myxoma had been made on the basis of echocardiographic findings alone: they had not undergone tumor resection for histopathologic study. Another 9 patients were excluded because their myxomas were confined to the right atrium. We thoroughly reviewed the medical records of the remaining 78 patients. Our review was approved by our institutional committee of human research.

The following data were collected: each patient's age, sex, clinical presentation (including cardiac, embolic, and constitutional symptoms), the time from the first onset of symptoms to diagnosis, characteristics of the cardiac myxoma, perioperative death, and morbidity. In the embolism group, findings of neurologic imaging (computed tomographic or magnetic resonance imaging of the brain, or magnetic resonance angiography) were recorded as well.

There were 48 females (62%; age range, 30–65 yrs) and 30 males (38%; age range, 37–65 yrs). Cardiac symptoms (including dyspnea, palpitation, chest pain or discomfort, and syncope) were noted in 55 patients (71%), embolic symptoms (including headache, hemiparesis, ataxia, aphasia, and memory impairment) in 15 (19%), and constitutional symptoms (including fatigue, weight loss, fever, and myalgia) in 14 (18%). In accordance with the symptoms and neurologic imaging findings (as described below), study patients were classified into 2 groups: embolism (15 patients, 19%) and nonembolism (63 patients, 81%).

All patients were monitored in the outpatient center, with clinical visits and echocardiography performed on an annual basis during the first 3 years after surgery. After 3 years, patients were monitored irregularly by mail or telephone contact. The follow-up was complete. The mean time to the last follow-up contact was 45± 34 months (range, 4–156 mo).

Results

Preoperative Echocardiography. Transthoracic echocardiography indicated a diagnosis of myxoma in every patient and was a simple, noninvasive method of morphologic examination of the cardiac chambers. Transesophageal echocardiography was also performed in 38 patients. The mean size of the cardiac myxomas was 3 ±3 cm (range, 1–7 cm). Most cardiac myxomas (75, or 96%) arose around the margin of the septal fossa ovalis, 2 (3%) arose from the posterior left atrial wall, and 1 (1%) arose from the base of the left atrium. The mean left ventricular ejection fraction (LVEF) was 0.56 ± 0.20 (range, 0.50–0.68). In addition to these findings, mild-to-moderate pulmonary hypertension was found in 22 patients, mild-to-moderate mitral insufficiency in 31 patients, and severe tricuspid insufficiency in 1 patient.

Neurologic Imaging. All 15 patients in the embolism group underwent cranial computed tomography; in addition, 10 underwent cranial magnetic resonance imaging and 5, magnetic resonance angiography. The neurologic imaging findings were consistent with ischemic cerebral infarction in 12 patients, but no intracerebral hemorrhage was observed. In the remaining 3 patients, no abnormality was detected by neurologic imaging. Transient ischemic attack was diagnosed by neurologists on the basis of characteristic symptoms and the neurologic imaging findings.

Surgical Resection. Surgical excision of tumors was performed through median sternotomy with cardiopulmonary bypass. The mean aortic cross-clamping time was 18 minutes (range, 15–48 min). Concomitant surgery was performed in 3 patients (4%): coronary artery bypass grafting in 2, and tricuspid valve repair in 1.

The myxomas were divided into 2 types, in accordance with the criteria described in previous clinicopathologic studies^4,5: type 1, with an irregular or villous surface and a soft consistency; and type 2, with a smooth surface and a compact consistency (Figs. 1 and 2). There were 42 patients with type 1 myxoma and 36 with type 2.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Comparison between the Embolism and Nonembolism Groups

We compared the clinical features of the embolism and nonembolism groups (Table I). The duration of symptoms was significantly greater in the embolism group than in the nonembolism group. Type 1 myxoma was more frequently noted in the embolism group than was type 2. The difference was significant (P=0.04).

TABLE I. Comparisons between the Embolism and Nonembolism Groups

View Fullscreen

Discussion

In the literature, neurologic manifestations have been reported in 10% to 30% of left atrial myxoma patients, with embolic cerebral infarct the most frequently observed event.^1,2,6 According to previous studies, neurologic and embolic events were more often associated with type 1 myxoma.^7,8 This type generally has a friable or gelatinous texture, which gives rise to emboli by erosion and fragmentation of the neoplasm itself (cellular, stromal, or both), or by erosion of thrombus covering the surface. Type 2 myxoma, solid and ovoid, appears firm and smooth and usually can be removed in 1 piece during operation. Emboli from type 2 myxomas, although few in number, can result from cracking of the tumor's surface by turbulent blood flow.^4,7,8

In 2 studies of central nervous system manifestations in cardiac myxoma, S.J. Lee⁷ and V.H. Lee⁸ and their respective colleagues showed that patients with cardiac myxoma who develop neurologic complications can lack concomitant cardiac symptoms. Accordingly, neurologic symptoms can present as the initial and only clinical manifestation. This subset of patients usually will be treated at onset by physicians other than cardiologists, which can delay diagnosis of myxoma and increase the risk of complications. Indeed this problem explains why, in our series, the time from onset of symptoms to diagnosis was much longer in the embolism group than in the nonembolism group.

The presence of cardiac symptoms actually directs diagnostic procedures to heart-imaging techniques, such as cardiac ultrasonography, and thereby accelerates diagnosis in this subset of patients. Although it has been suggested that some clinical examinations, such as auscultation, imply the presence of cardiac myxoma in a patient who has only neurologic manifestations, no series has yet confirmed this hypothesis. Neither could we, in our group, identify cardiac myxoma by auscultation. Whatever the case, if embolic stroke is suspected on the basis of neurologic symptoms, the possibility of cardiac myxoma must be borne in mind.^4,7,8

Another interesting finding of our study is that recurrent cerebral infarction before surgical removal of tumor is not rare. One patient had 4 episodes of cerebral infarction during a period of 2 years, another had 2 in 8 months, and a 3rd patient had 2 in 6 months. In the embolism group, 4 patients had at least 1 onset of TIA before arrival at our institute with a final diagnosis of cardiac myxoma. Although peripheral embolism has occasionally been reported,^9–11 it was not observed in our series.

Limitations. Our study has 3 major limitations. First, it was a retrospective study in a single institution. Second, neurologic imaging was applied only to patients with an embolic event as the initial manifestation, rather than to all patients. Consequently, patients who experienced asymptomatic cerebral infarctions would have been classified into the nonembolism group. Third, we could not directly confirm the relationship between cerebral embolism and cardiac myxoma, because we could not find any tumor-surface thrombus that could be related to embolism.