A 31-year-old man with a long history of paroxysmal atrial fibrillation underwent uneventful transcatheter radiofrequency ablation of an ectopic electrical focus in the left superior pulmonary vein. The patient had been taking oral anticoagulants for the previous 2 months. Unfortunately, atrial fibrillation recurred 3 days after the procedure, and the anticoagulation regimen was reinstituted. One month after the procedure, the patient began to experience daily hemoptysis (3-5 mL of blood) and episodes of transient left chest discomfort. He presented for evaluation 4 months after the procedure.

At physical examination, the patient had a regular heart rate and rhythm, and his lungs were clear to auscultation. All laboratory values were within the expected range except a prothrombin time of 16.9 sec, attributable to the patient's regimen of oral warfarin. Chest radiography showed a poorly marginated left upper lobe opacity, suggestive of asymmetric pulmonary edema. The patient was afebrile, and infection was considered unlikely. Contrast-enhanced CT revealed an occluded left superior pulmonary vein, increased attenuation in the mediastinum adjacent to the vein, poorly marginated nodules and diffuse ground-glass opacities in the left upper lobe, aortopulmonary window adenopathy, and pleural thickening in the left hemithorax (Fig. 1A,1B,1C,1D). Occlusion of the left superior pulmonary vein was confirmed at cardiac catheterization. Ventilation—perfusion scintigraphy showed an absence of perfusion of the left upper lobe.

Because of venous occlusion and recurrent hemoptysis, the patient was taken to the operating room and underwent a left upper lobectomy. At surgery, he was found to have a markedly scarred left upper lobe pulmonary vein. A thick pleural peel was present, and decortication was also performed. The patient tolerated the procedure well and was subsequently discharged. Histopathologic examination of the resected left upper lobe showed venous occlusion and hemorrhagic infarction with associated hypertensive arteriopathy. Lymph nodes removed at surgery showed no evidence of malignancy, granulomatous inflammation, or fungal or bacterial organisms.

Transcatheter radiofrequency ablation is a new treatment for atrial fibrillation. Over 90% of ectopic beats that lead to atrial fibrillation originate in the pulmonary veins, almost half in the left superior pulmonary vein [1]. In most individuals, the sleeves of left atrial myocardium extend into the pulmonary veins for a distance that varies from 2 to 17 mm [2] and may cause ectopic electrical activity. Treatment involves mapping and provoking these ectopic foci, and then performing radiofrequency ablation using up to 50 W of power. In the largest series, this technique successfully eliminated atrial fibrillation in more than 70% of patients and provided symptomatic relief in the remainder [3]. The major benefit of the procedure is that it allows antiarrhythmia pharmacologic and oral anticoagulation therapies to be discontinued.

The types and rates of complications associated with radiofrequency ablation for treatment of atrial fibrillation have yet to be fully elucidated. An immediate burning sensation, cough, or tachycardia are frequently elicited in patients during the procedure. After the procedure, more than 40% of patients develop mild pulmonary vein stenosis, presumably due to thermal injury. Severe stenoses may result from more extensive ablation, particularly in patients in whom more than 30 W of power are used [3]. Local irritation of the vascular endothelium and luminal narrowing may result in pulmonary venous hypertension and, in cases in which the vein thromboses, venous infarction.

To our knowledge, only one other case of postablation venoocclusive disease has been described [4]. In that patient, severe stenosis after ablation of the left superior pulmonary vein was discovered 10 days after the procedure and was successfully managed using balloon dilation [4]. In our case, however, the patient was treated by left upper lobectomy for several reasons: the long time period between the procedure and identification of pulmonary vein stenosis (4 months), evidence of complete occlusion of the vein on CT and at cardiac catheterization, and complete absence of left upper lobe perfusion at ^99mTc-macroaggregated albumin ventilation—perfusion scintigraphy.

The diagnosis of pulmonary vein stenosis after radiofrequency ablation has been made by revealing an increase in venous flow velocity on transesophageal echocardiography or by directly showing venous narrowing at cardiac catheterization or on MR imaging [5, 6]. The diameter of the pulmonary vein ostia as measured at cardiac catheterization or anatomic dissection are quite variable [2, 7], ranging from 9 to 13 mm for the left superior pulmonary vein; from 8 to 12 mm for the right superior pulmonary vein; from 3 to 9 mm for the left inferior pulmonary vein; and from 3 to 7 mm for the right inferior pulmonary vein. These measurements provide a guideline when one is interpreting CT and MR imaging examinations to identify possible pulmonary venous stenosis, but normal ranges for the pulmonary vein diameter at cross-sectional imaging have not been described.

In our patient, contrast-enhanced CT was not only useful for revealing direct evidence of pulmonary venous occlusion but also for showing indirect evidence of physiologic obstruction. These included findings of septal thickening in the left upper lobe indicative of localized pulmonary venous hypertension and peripheral opacities consistent with the venous infarction confirmed at histopathologic examination. These findings are similar to those described in cases of idiopathic pulmonary venoocclusive disease [8]. We suspect, on the basis of this case and our experience imaging other patients with possible pulmonary vein stenosis after radiofrequency ablation, that although mild degrees of stenosis may be common and difficult to diagnose because of normal variation in vein diameter, hemodynamically important stenoses are likely to be accompanied by the parenchymal findings we described. In other words, the presence or absence of appropriate parenchymal findings may be more important than the finding of mild pulmonary venous narrowing on contrast-enhanced CT.

In addition to direct and indirect evidence of pulmonary venous occlusion, we also saw increased attenuation in the mediastinal fat adjacent to the left superior pulmonary vein in our patient, as well as localized lymphadenopathy. These findings suggest that the thermal injury caused by radiofrequency ablation led to localized mediastinal inflammation and fibrosis, probably compounding the direct venous injury. However, because of the associated mediastinal findings, other causes of pulmonary vein stenosis were considered in the differential diagnosis. Idiopathic fibrosing mediastinitis was considered but thought unlikely because it usually manifests on CT as a mass of diffusely infiltrating soft-tissue attenuation involving multiple mediastinal compartments. Fibrosing mediastinitis due to histoplasmosis can manifest as a focal process, as in our patient, but often appears as extensive calcification in the mass. Lung cancer was also considered because it can manifest as a focal mass obstructing the pulmonary vein. However, the age of our patient and the association of symptoms and radiologic findings with the history of radiofrequency ablation militated against the diagnosis of either cancer or histoplasmosis, and at surgery and histopathologic examination, no evidence of either process was found.

Address correspondence to J. G. Ravenel.