Contrast-Enhanced CT Scans at the Level of the Right Atrium.

(Figure 1Figure 1

Contrast-Enhanced CT Scans at the Level of the Right Atrium.).

The finding of a right atrial density adjacent to the infusion port suggests a source for pulmonary embolism. It most likely represents a thrombus, although a tumor, either metastatic breast carcinoma or atrial myxoma, is also possible. The severe hypoxemia can be explained by recurrent microemboli. Ideally, the infusion port should be removed, since it is probably the nidus of thrombus formation. Because the thrombus appears to be large, thrombolytic therapy may be indicated to prevent massive embolization or obstruction of right ventricular outflow before removal of the infusion port, but there is a risk that removal of the port could dislodge the thrombus and precipitate embolization. An inferior vena caval filter would not provide protection against such an event. There is an increased risk of hemorrhage in a patient with cancer; brain imaging to rule out metastases should be considered. For now, I would favor treatment with heparin to stabilize the clot. Because the right atrial mass was not seen on a transthoracic echocardiogram, I would request a transesophageal echocardiogram, which provides better visualization of the atria.

 

A transesophageal echocardiogram was obtained. It confirmed the presence of a large, mobile thrombus in the right atrium, arising from the tip of the central venous catheter. The thrombus occupied a large portion of the right atrium and abutted the tricuspid orifice Figure 2

Transesophageal Echocardiogram.

(Figure 2Figure 2

Transesophageal Echocardiogram.; and Video 1, available with the full text of this article at

NEJM.org). There was also a moderate-size patent foramen ovale with bowing of the right atrial septum, which was consistent with elevated right atrial pressure (Video 1).

 

I am very concerned about the risk of embolization from a large thrombus in the right atrium, particularly given the moderate-size patent foramen ovale, which presents a risk of systemic embolization. Catheter-delivered thrombolysis is my preferred option, although it clearly has risks in this patient, and I remain concerned that it could trigger embolization or bleeding. I would first perform magnetic resonance imaging of the brain to look for a metastatic tumor, which if present, would contraindicate thrombolytic therapy.

The risk of precipitating pulmonary or systemic embolization with thrombolysis was deemed too high. An interventional radiologist and a thoracic surgeon were reluctant to directly remove the thrombus. A decision was made to continue low-molecular-weight heparin and reassess the thrombus after 48 hours. During this time, the patient became progressively more hypoxic, despite an increase in supplemental oxygen to 100%. The nurses noted that the patient's hypoxemia was substantially worse when she was lying on her left side.

Hypoxemia that is resistant to supplemental oxygen suggests intrapulmonary or intracardiac shunting. Positional hypoxemia is typically seen with localized or lateralizing pulmonary disease, because the shunt fraction increases when the abnormal lung is dependent. No pulmonary disease has been identified in this patient, and thus I am concerned about the possibility of right-to-left blood flow through the patent foramen ovale. Such flow can be confirmed by means of echocardiography performed after the administration of agitated saline, especially if the study is performed when the patient is lying in the decubitus position. Pending further evaluation, this patient should be positioned on her right side.

Ventilation–perfusion scanning showed no perfusion defects. It revealed immediate uptake Figure 3

Scintigraphic Lung-Perfusion Scan.

Of radioisotope in the kidneys, consistent with a right-to-left shunt (Figure 3Figure 3

Scintigraphic Lung-Perfusion Scan.). To confirm the

Presence of a shunt through the patent foramen ovale, saline-contrast echocardiography was performed with the patient lying in the left and right decubitus positions. Administration of the contrast agent through the antecubital vein resulted in immediate opacification of the chambers on the left side in all positions. The rate and extent of opacification were increased when the patient was lying on her left side (Video 2).

These investigations have confirmed the presence of a right-to-left shunt through a patent foramen ovale. This is the likely cause of the hypoxemia. The thrombus is partially obstructing the tricuspid valve. I suspect that this may have increased right atrial pressure and caused the right-to-left shunt. The position of the thrombus may also be redirecting right atrial flow toward the patent foramen ovale. Either mechanism could be accentuated when the patient is lying on her left side, because the location of the thrombus within the atrium changes with the change in position. Again, there is a risk of systemic paradoxical embolization and refractory hypoxemia. At this stage, removal of the central venous catheter and the thrombus and closure of the patent foramen ovale would be the optimal management strategy.

The cardiothoracic-surgery service recommended removal of the central venous catheter with the thrombus and closure of the patent foramen ovale by means of sternotomy. At surgery, a 3.2 cm by 2.8 cm thrombus was removed. It was adherent to the catheter and almost completely occupied the right atrium. The defect in the atrial wall was repaired. The patient was extubated immediately, and 4 hours postoperatively, her oxygen saturation was 99% while she was receiving oxygen at a rate of 2 liters per minute. She was discharged on the fifth postoperative day; she did not require supplemental oxygen, and her oxygen saturation was in the normal range in all positions.

Commentary

 

Clinical problem-solving requires not only the interpretation of results of multiple tests but also often the reinterpretation of these tests as the clinical case evolves. The onset of respiratory failure in association with cancer treatment has a wide differential diagnosis. In this case, pulmonary embolism was recognized as the major concern, yet the negative findings on CT pulmonary angiography were initially interpreted as ruling out this diagnosis. The discussant, however, recognized the importance of interpreting the CT pulmonary angiographic results in the context of the clinical probability of pulmonary embolism, which was considered high in this patient with cancer and a central venous catheter.

 

In cases in which the pretest suspicion of pulmonary embolism is moderate or high, the risk of a false negative result on CT pulmonary angiography is considerable (the rates in one study were 11% and 40%, respectively).1 Long-term use of a central venous catheter is complicated by venous thrombosis in up to 63% of cases and by pulmonary embolism in 4% of cases.2 These complications are often not suspected on clinical examination. Major risk factors for central-venous-catheter thrombosis include placement on the left side; placement of the tip in the upper superior vena cava; the presence of cancer, particularly adenocarcinoma; and the use of sclerosing chemotherapy. 2 Appreciating the discordance between the clinical risk of pulmonary embolism and the results of CT pulmonary angiography, the discussant recommended further testing. This did not identify a pulmonary embolism, but it did identify a central-venous-catheter thrombus. In this context, one would have expected to find elevated jugular venous pressure and a gradient across the right ventricular inflow on echocardiographic examination. Their absence may be explained by the positional nature of the disease and by the technical difficulties inherent in performing echocardiography in an acutely ill patient.

 

When a central-venous-catheter thrombus is identified, removal of the catheter is generally recommended, although data from randomized trials are lacking to guide management. The size of the thrombus and the identification of a patent foramen ovale on echocardiography raised concern about systemic embolism. The foramen ovale remains patent in 25% of the population and warrants routine consideration in cases in which right-to-left shunting is suspected.3 Although patent foramen ovale has been associated with cryptogenic stroke, the available literature suggests that the risk of clinically significant systemic embolism is low in patients who have a patent foramen ovale and central-venous-catheter thrombosis.4

 

The nurses' recognition that the hypoxemia was more pronounced when the patient was in the left lateral position was a crucial piece of information in this case. The discussant recognized that hypoxemia due to a right-to-left shunt through a patent foramen ovale can be positional. This most typically results in the platypnea–orthodeoxia syndrome, a rare disorder characterized by dyspnea and hypoxemia when the patient is in the upright, but not the supine, position.5 In this case, hypoxemia was exaggerated in the left lateral position. The occurrence of dyspnea when the patient is lying on one side, but not the other, is termed trepopnea. It usually occurs when there is an abnormality in only one lung, resulting in a lateralized ventilation–perfusion mismatch or shunt.6 Moving the diseased lung into a dependent position increases blood flow to it, which results in increased shunt and hypoxemia. Trepopnea has also been described in association with atrial defects, including patent foramen ovale. 7 Recognition of the phenomenon in this case led to further testing with a lung-perfusion scan, which ruled out pulmonary embolism (normal lung perfusion) but also showed early uptake in the kidneys, which was consistent with a right-to-left shunt; a subsequent saline-contrast echocardiogram confirmed an intraatrial shunt.

 

An intraatrial right-to-left shunt requires an anatomical defect, such as an atrial septal defect or patent foramen ovale, and an alteration in the normal right atrial hemodynamics. Normal right atrial pressure and right ventricular compliance favor left-to-right shunting. Elevation in right atrial pressure can reverse the direction of flow, causing right-to-left shunting and hypoxemia. This is most commonly associated with pulmonary hypertension, but it can also occur with tricuspid stenosis or mechanical obstruction of the tricuspid valve by an atrial myxoma or right atrial thrombus.8,9 In these cases, the pathophysiology reflects elevated right atrial pressure despite normal pulmonary pressures and right ventricular compliance. Furthermore, lying in the lateral position can redirect blood flow toward the atrial defect rather than the tricuspid valve, increasing right atrial pressure. This may explain the finding of trepopnea in atrial defects, as outlined above. In the present case, we speculate that the acute pathophysiological event was the formation of a large thrombus at the distal end of the central venous catheter, creating functional tricuspid stenosis, as evidenced by elevated right atrial pressure in association with normal right ventricular pressure. This presumably led to a right-to-left shunt through the patent foramen ovale, resulting in severe hypoxemia.

 

The discussant initially considered treatment with catheter-delivered thrombolysis. Data from randomized, controlled trials are lacking to compare therapies, but a retrospective observational study of patients with right heart thromboemboli showed a higher survival rate among those treated with thrombolysis than among those treated with anticoagulation alone or thrombectomy.10 However, thrombolysis does not remove the initiating focus, and it carries risks of major hemorrhage and massive embolism, with the degree of risk related to the relative mobility of the thrombus on the echocardiogram.11 The presence of a right-to-left shunt increases the risk of systemic embolism. Whereas percutaneous closure of atrial defects is now common practice for patients at risk for paradoxical embolization, the presence of an atrial thrombus is generally considered a contraindication.12 In view of this patient's clinical deterioration, surgical thrombectomy with repair of the atrial defect was ultimately considered the preferred option.13

 

This case reinforces the importance of interpreting test results with caution and with attention to the clinical picture. Effective communication between the nurses and specialists caring for the patient led to the correct diagnosis and a good clinical outcome. This clinical scenario and clinical problem-solving exercise emphasize the importance of looking at the “hole” picture.