Cardiovascular Collapse Induced by Position-dependent Pulmonary Vein Occlusion in a Patient with Fibrosing Mediastinitis

A 27-yr-old woman was admitted for evaluation of severe progressive dyspnea on exertion and a nonproductive cough. Soon after development of these symptoms, 5 months before admission to the Mayo Clinic (Rochester, Minnesota), her chest radiogram revealed a right lower lobe infiltrate, which worsened despite therapy with levofloxacin, and hemoptysis developed. A repeat chest radiogram and computed tomography scan of the chest, 3 months after initial presentation, demonstrated bilateral ground glass infiltrates with diffuse mediastinal lymphadenopathy. All infectious and rheumatologic test results were negative, whereas pulmonary function tests revealed a restrictive pattern. Transthoracic echocardiography indicated pulmonary hypertension with significantly reduced right ventricular function (cor pulmonale). The patient underwent a mediastinoscopy, but pathology was unrevealing. Mediastinoscopy was complicated by sudden “oxyhemoglobin desaturation and cardiovascular collapse” requiring mechanical ventilation and inotropic support. The patient was stabilized in the intensive care unit, and in the absence of explanation for cardiovascular collapse, she was transferred to the Mayo Clinic for further evaluation.

At admission, the patient was stable, requiring no additional oxygen, and showed no signs of respiratory distress. Immediate bronchoscopy was scheduled. She arrived to the bronchoscopy suite sitting comfortably in her stretcher. After standard American Society of Anesthesiologists monitors were applied and after blood pressure (100/70 mmHg), heart rate (90 beats/min), and oxyhemoglobin saturation (Spo²91% on room air) were recorded, the pulmonologist used lidocaine to anesthetize the upper airway with the patient in the semisitting position. The patient tolerated awake tracheal intubation with 1 mg midazolam. After her trachea was intubated, she was positioned supine for flexible bronchoscopy. After initiation of bronchoscopy, the patient became “uncooperative”; her oxyhemoglobin saturation decreased to the low 80s, and her blood pressure decreased to 65/35 mmHg. Bronchoscopy was aborted, and pure oxygen was administered via  the endotracheal tube. At this time, the consultant anesthesiologist was urgently summoned. In addition to fluid boluses, several 100-μg doses of phenylephrine were given, and with each bolus, the patient’s arterial blood pressure increased and respiratory distress lessened. While the anesthesiologist was resuscitating the patient, the pulmonologist informed him of this patient’s underlying mediastinal pathology as well that there was no bronchial compression visualized during bronchoscopy. Because no resistance was encountered during bag ventilation, an assumption was made that the patient’s problem was arising from some type of mechanical compression on pulmonary vasculature. Therefore, to facilitate pulmonary blood flow, we did not administer muscle relaxants, and we maintained the patient breathing spontaneously. In addition, we positioned the patient in steep reverse-Trendelenburg posture with the back further elevated to near sitting position. With this maneuvering, hemodynamics and oxyhemoglobin saturation normalized. An infusion of propofol at 40 μg · kg⁻¹· min⁻¹was initiated. At that point, we inserted a brachial arterial catheter; arterial blood pressure was 85/50 mmHg. An infusion of phenylephrine was initiated to keep the systolic blood pressure above 90 mmHg for transfer to the intensive care unit. During transport, the patient’s oxyhemoglobin saturation was in low 90s, while spontaneous breathing was assisted with bag ventilation. At arrival to the intensive care unit, a pulmonary artery catheter was inserted; cardiac index was 2.3 l · min⁻¹· m⁻², pulmonary artery occlusion pressure was 38 mmHg, and pulmonary artery pressure was 86/47 mmHg (at systemic arterial blood pressure of 110/85 mmHg). Epoprostenol (970 ng/min) as well as nitric oxide was added, with minimal effects on pulmonary artery pressure (68/41 mmHg). Soon after, the patient’s arterial blood pressure decreased to 59/31 mmHg, and her oxyhemoglobin saturation decreased to 73%. At that time, we initiated controlled ventilation with 100% oxygen. To maintain systolic arterial blood pressure above 100 mmHg, high-dose norepinephrine and dobutamine were infused together with additional fluid boluses. A transthoracic echocardiogram showed right ventricular enlargement with a moderate decrease in function and estimated right ventricular systolic pressure of 69 mmHg. The left atrium was of normal size, the left ventricular ejection fraction was 65%, and the patient had no valvular (mitral) disease. High pulmonary artery occlusion pressure in context of normal left ventricular function and anatomy (no mitral stenosis) suggested the possibility of pulmonary vein occlusion as a cause for pulmonary hypertension (postcapillary pulmonary hypertension). The transesophageal echocardiogram confirmed high-flow velocity in the single pulmonary vein that was able to be visualized, consistent with a narrowed vein lumen. A computed tomography scan (fig. 1) and computed tomography angiogram confirmed near-total occlusion of the right inferior and left superior pulmonary veins along with moderate occlusion of the right superior pulmonary vein; only the left inferior pulmonary vein was patent. The patient’s condition continued to deteriorate rapidly, and she was unable to maintain blood pressure and oxyhemoglobin saturation while supine, even with vasopressors. It was repeatedly noted that the patient’s oxyhemoglobin saturation and hemodynamics fared much better when she was turned on her side, suggesting position-dependent occlusion of her pulmonary vessels. An interventional cardiologist was consulted, and the patient was emergently brought to the catheterization laboratory, where successful angioplasty of the right superior pulmonary vein via  transseptal puncture was performed. Given her high fever (41.6 °C) suggestive of infection, the stents were not used. After single-vessel angioplasty, the patient’s hemodynamics improved, and at that point, it was believed that the patient’s “state of infection” should be brought under control before performing more extensive procedures, including stenting. All blood cultures test results remained negative, but fever persisted despite wide-spectrum empiric antibiotic coverage. Serology test results were positive for previous exposure to histoplasmosis. While the patient was in the intensive care unit, her clinical status continued to deteriorate rapidly, and she died 3 days after angioplasty.

Autopsy confirmed the diagnosis of sclerosing mediastinitis producing bilateral pulmonary hilar encasement, extrinsic pulmonary vein compression, and severe pulmonary congestion. A fibrous mediastinal mass with necrosis in the middle (explaining the fever) was partially encasing the left and right pulmonary hilar regions constricting the vasculature. Fungal stains showed rare organisms consistent with Histoplasma  species.

Large anterior mediastinal masses can cause compression of mediastinal structures, causing airway obstruction or cardiovascular compromise.10The additive effects of anesthesia, paralysis, and supine positioning can lead to acute airway or vascular obstruction and death.10,11These respiratory and hemodynamic problems may resolve in the sitting, prone, semiprone, or lateral decubitus position.9,10Contrary to anterior mediastinal pathology, it is generally considered that diseases in the middle mediastinal compartment do not have significant implications for anesthetic management. However, our case illustrates that fibrosing mediastinitis, a disease distributed mostly throughout the middle mediastinal compartment, can cause equally severe hemodynamic and oxygenation compromise, resembling the clinical picture of an anterior mediastinal mass. The fibrotic process in patients with advanced fibrosing mediastinitis may affect mediastinal structures to produce airway or vessel occlusion, most frequently causing superior vena cava syndrome.12,13Pulmonary artery or vein compressions are less frequently described.1,7,14To the best of our knowledge, position-dependent pulmonary vein compression, as present in our patient, has never been reported.

Pulmonary hypertension and cor pulmonale associated with fibrosing mediastinitis are caused by critical obstruction of pulmonary venous outflow.15,16This is potentially a reversible cause of pulmonary hypertension, because relieving the obstruction with angioplasty normalizes venous inflow into the left atrium. Pulmonary artery catheter pressure measurements may provide a significant clue in elucidating the mechanism of pulmonary hypertension, and when combined with an echocardiogram, the mechanism of pulmonary hypertension may be even further pinpointed. In fact, high pulmonary artery end-diastolic pressure (47 mmHg in our patient) which is 5–15 mmHg higher than pulmonary artery occlusion pressure (38 mm in our patient) strongly suggested postcapillary pulmonary hypertension in our patient, i.e. , pulmonary venoocclusive disease.17Postcapillary pulmonary hypertension is caused by either left heart–related factors (failure, mitral stenosis, atrial myxoma) or pulmonary venous factors (pulmonary vein stenosis or other type of pulmonary venoocclusive disease). Because our patient’s echocardiography demonstrated normal left ventricular function and anatomy (no valvular disease, thrombus, etc. ), the most plausible etiology of postcapillary pulmonary hypertension was venous obstruction, associated with mediastinal disease. This hypertension may be subject to correction with stenting; therefore, it is important to make an early diagnosis before the disease advances and irreversible parenchymal changes occur.1,16Furthermore, sudden onset of pulmonary hypertension in conjunction with mediastinal pathology, in a previously healthy young patient, should have raised a high suspicion of mechanical cause, such as venous outflow obstruction, much earlier in the course of her disease.

Radiographic abnormalities in patients with fibrosing mediastinitis consist of diffuse lesions, atelectasis, and infiltrates. Chest radiographs may occasionally be deceptively unremarkable, even in the presence of major central airway or vascular occlusion, especially when the lesion is subcarinal.18Our patient had pathology in the middle mediastinum, which radiologically, from an anesthetic (mechanical) standpoint, seemed benign, but at the same time, total occlusion of three pulmonary veins was present. Furthermore, our case is unique in the sense that the pulmonary vein occlusion was posture dependent, the same as in patients with anterior mediastinal masses.9Unfortunately, an episode of “cardiovascular collapse” during mediastinoscopy at an outside institution was only retroactively related to this patient’s pulmonary vein pathology.

In conclusion, we describe a patient with mediastinal fibrosis and position-dependent occlusion of pulmonary veins. Her presenting chest radiogram was deceiving, and providers in two institutions were puzzled with hemodynamic and oxygenation problems they encountered after tracheal intubation with the patient in supine posture. Our case illustrates that if intractable hypotension after induction of anesthesia occurs, even in patients with midmediastinal pathology, a possibility of vascular compression should be considered. Finally, acute development of pulmonary hypertension in previously healthy young individuals, associated with radiologic changes in the mediastinum, should alert physicians to search for venoocclusive etiology of pulmonary hypertension. Information obtained from pulmonary catheter measurements in conjunction with transthoracic echocardiographic findings can give us important clues regarding the mechanism of pulmonary hypertension.