The Prevalence of Deep Venous Thrombosis in Patients with Suspected Paradoxical Embolism

According to a recent survey [1], paradoxical embolism is considered an infrequent event, accounting for fewer than 2% of all arterial emboli. Diagnosis of paradoxical embolism is based on three criteria: evidence of arterial embolism, a right-to-left shunt, and a thrombosis in the venous system or in the right atrium. Detection of small intracardiac shunts like those in patients with a patent foramen ovale has been facilitated by contrast echocardiography. Use of this technique has resulted in the identification of young patients with embolic stroke presumably caused by paradoxical embolism via a patent foramen ovale [1-3]. In a recent study, patients with cryptogenic stroke had a higher prevalence of patent foramen ovale than patients with stroke of determined cause in all age groups, even after adjustment for the presence of recognized risk factors for stroke [4]. However, investigators in previous studies did not do transesophageal echocardiographic evaluations for alternative cardiogenic causes or assess the frequency of venous thrombosis. The objectives of our study were 1) to assess the frequency of venous thrombosis in patients with suspected paradoxical embolism in whom a patent foramen ovale detected by transesophageal echocardiography was the only possible cause of embolism; and 2) to assess the effect of a history of venous thromboembolism or arterial embolic events. Methods Patients Between December 1988 and January 1991, 406 patients with clinically suspected embolic events underwent transesophageal echocardiographic evaluation to search for a cardiac cause. Contrast studies were done in the 264 patients in whom established sources of emboli (endocarditic vegetations, mitral valve stenosis, left ventricular thrombi or dyssynergy, spontaneous echocardiographic contrast, left atrial or appendage thrombi, or a myxoma) could not be found. Forty-nine patients found to have a right-to-left shunt at the atrial level constituted the study sample. They were referred to the echocardiographic laboratory from various medical and surgical departments as well as from the neurology department. Physical examination, routine laboratory tests, electrocardiographic study, and chest radiography were done in all patients. No patient showed clinical evidence of an abnormal communication between the systemic and the pulmonary circulation. Diagnosis of Embolism All patients with limb ischemia underwent angiography. Brain computed tomographic scanning and duplex sonographic studies of the carotid arteries were done in all patients with stroke. Cerebral embolism was diagnosed in patients if a computed tomographic scan showed a territorial infarction corresponding to the neurologic deficit and if arteriosclerotic lesions of the ipsilateral carotid artery were excluded on the basis of the sonographic studies [5]. Anticoagulant Treatment All patients with stroke received antithrombotic prophylaxis starting on the day of admission with either 5000 units of low-molecular-weight heparin (Fragmin, Kabi Pharmacia, Uppsala, Sweden) administered once daily or 5000 units of unfractionated heparin (Liquemin, Hoffmann-LaRoche AG, Basel, Switzerland) administered subcutaneously 2 or 3 times daily. In patients with embolic occlusion of a limb artery or documented venous thrombosis, heparin was infused continuously at a dosage that prolonged the activated partial thromboplastin time from 50 to 70 seconds. After 1 week of this treatment, these patients were usually given oral anticoagulant therapy. Diagnosis of Patent Foramen Ovale All patients underwent transthoracic and transesophageal echocardiographic studies after informed consent had been obtained. Either a Diasonics CV (Diasonics Inc., Milpitas, California) or an Aloka SSD 870 (Aloka Co., Tokyo, Japan) with a 3.5-mHz transducer was used for transthoracic echocardiography, and either a Diasonics Varian 3400 with a monoplane 3.5-mHz probe or an Aloka SSD 870 with a biplane 5-mHz probe was used for the transesophageal echocardiography. After visualizing both atria and the interatrial septum with the transesophageal probe, 10 mL of agitated normal saline solution was injected into an antecubital vein while the patient performed the Valsalva maneuver. When the first microbubbles appeared in the right atrium, the patient was asked to resume normal breathing. Contrast injections were repeated during normal respiration and, in cases of doubt, during a second Valsalva maneuver after a 3-minute interval to allow for disappearance of all microbubbles. A patent foramen ovale was diagnosed if two or more microbubbles crossed the interatrial septum immediately after the appearance of contrast material in the right atrium [6]. The transesophageal studies were recorded on videotapes and reviewed by a second observer from the same laboratory who was blinded to the initial assessment. Agreement on presence or absence of patent foramen ovale was reached in all cases. Clinical Assessment Immediately after the transesophageal echocardiographic study, patients with a patent foramen ovale were asked by the physician who did the examination about a history of previous venous thrombosis or pulmonary or systemic embolism and about a temporal relation of the actual embolic event with a Valsalva maneuver. Thereafter, the same physician inspected the legs for signs of venous thrombosis. All findings were recorded for later analysis. Screening for pulmonary embolism was not done routinely, but patients with clinical signs underwent ventilation-perfusion scanning. Diagnosis of Venous Thrombosis Bilateral ascending venography was proposed to all patients with a patent foramen ovale, and informed consent was obtained. The contrast agent used was iopamidol (Jopamiro, Gerot Pharmazeutika, Vienna, Austria [licensed by Bracco S.p.A., Milano, Italy]). Venous thrombosis was diagnosed in the presence of a constant intraluminal filling defect assessed by two independent observers. The study design was approved by our institutional review board. For both ethical and methodologic reasons, we decided not to recruit a control group for venographic evaluation. Because of the heterogeneity of the study sample, the sequence of individual examinations and intervals between examinations differed among patients. Statistical Analysis Data were analyzed using the Student t-test for numerical variables and the Fisher exact test for categorical variables (SAS Institute Inc., Cary, North Carolina). All P values are two-tailed. A P value less than 0.05 was considered to indicate significance [7]. The 95% CIs for rates were calculated using the appropriate tables in Documenta Geigy [8]. Results Characteristics of Patients with Patent Foramen Ovale Transesophageal contrast study revealed a patent foramen ovale in 49 of 264 patients (19%; CI, 15% to 25%). Of these 49 patients, 24 were women and 25 were men (mean age, 58 years; age range, 31 to 85 years). The 49 patients had been admitted for stroke (36 cases), limb ischemia (7 cases), and venographically confirmed venous thrombosis (6 cases). The 6 patients with venous thrombosis (3 women and 3 men [mean age, 65 years; age range, 38 to 85 years]) developed arterial embolism during their clinical stay; 5 experienced stroke and 1 had brachial artery embolism. Contrast transesophageal studies were done 6 to 24 hours after the embolic event and 1 to 4 days after confirmation of venous thrombosis by venographic study. None of the remaining 43 patients showed clinical evidence of venous thrombosis of the legs. However, assessment was hampered in 3 of these 43 patients by acute ischemia caused by embolism in the femoral artery. Pulmonary embolism was suspected in only 5 of the 49 patients. These 5 patients were admitted because of vein thrombosis, and ventilation-perfusion scan studies confirmed pulmonary embolism in all of them. Seven patients without clinical signs of venous thrombosis (one woman and six men [mean age, 57 years; age range, 31 to 77 years]) did not undergo venographic study for various reasons: An alternative mechanism for the index event had been detected in four patients (one each with brain tumor, lacunar stroke, carotid artery stenosis, and subclavian artery stenosis). Venography was declined by the referring physician in two cases because the patients were known to have adverse reactions to the contrast agent, and in one case the patient refused to undergo venographic study. Other Causes of Ischemic Events Thirteen of the 42 patients who underwent venographic study had potential alternative causes of their embolic event: Artery-to-artery embolism was the possible cause of ischemia in 4 patients (carotid artery stenosis [2 patients], and subclavian artery stenosis (1 patient), popliteal artery stenosis [1 patient]). Lacunar lesion was the possible cause of stroke in 4 other patients. Embolism from the left atrium had to be considered in 5 patients with chronic or intermittent atrial fibrillation, but neither left atrial thrombi nor spontaneous echocardiographic contrast was detected in these patients. Prevalence of Venous Thrombosis Venographic study was done in 42 patients and showed abnormal findings in 24 patients (Table 1). The procedure was well tolerated by all patients, and no complications due to the contrast agent were observed. Venographic study was done within 1 week before or after the embolic event in 17 patients and showed venous thrombosis in 15 patients (Table 2). The other 25 patients underwent venographic study 8 to 90 days after the embolic event, and venous thrombosis was detected in 9 patients, yielding a significantly lower rate than that seen in patients who underwent evaluation within a week of the embolic event (P = 0.001). Of the 42 patients who underwent venographic study, 29 patients (16 women and 13 men [mean age, 59 years; range, 37 to 85 years]) had no detectable alternative cause of their ischemic event, and the only explanation for their even