Abstract
Haemoptysis is a potentially life-threatening condition with the need for prompt diagnosis. In about 10–20% of all cases the bleeding source remains unexplained with the standard diagnostic approach. The aim of this article is to show the necessity of widening the diagnostic approach to haemoptysis with consideration of pulmonary venous stenosis as a possible cause of even severe haemoptysis and haemoptoe.
A review of the literature was performed using the Medline/PubMed database with the terms: “pulmonary venous stenosis”, “pulmonary venous infarction” and “haemoptysis”. Further references from the case reports were considered.
58 case reports and case collections about patients with haemoptysis due to pulmonary venous stenosis were detected. This review gives an overview about the case reports and discusses the underlying pathophysiology and the pros and cons of different imaging techniques for the detection of pulmonary venous stenosis.
Several conditions predispose to the obstruction of the mediastinal pulmonary veins. Clinical findings are unspecific and may be misleading. Pulmonary venous stenosis can be detected using several imaging techniques, yet three-dimensional magnetic resonance-angiography and three-dimensional contrast-enhanced computed tomography are the most appropriate. Pulmonary venous stenosis should be considered in patients with haemoptysis.
Introduction
This article gives an overview about pulmonary venous stenosis as a possible cause of haemoptysis. Inflammatory and malignant pulmonary diseases may cause haemoptysis, and diagnostic algorithms as management principles for haemoptysis are well described. However, with the standard diagnostic approach, including bronchoscopy, computed tomography (CT) and bronchial angiography, in about 10–20% of cases the bleeding source remains unexplained [1]. Although many cases were reported of patients with severe haemoptysis due to pulmonary venous obstruction [2–59] (online supplementary table S1), the pulmonary veins are only infrequently mentioned in the literature about the management of haemoptysis [60–65]. Besides the rare condition of congenital pulmonary venous stenosis, there are several clinical conditions predisposing to the obstruction of the central pulmonary veins, like mediastinal masses such as solid neoplasms or bulky lymphoma, fibrosing mediastinitis and mediastinal granulomatous diseases. In addition, lung transplantation [66, 67] and lobectomy or bilobectomy [24] may result in pulmonary venous stenosis (table 1). Furthermore, cor triatriatum [53], left atrial myxoma and post-pneumonectomy syndrome may cause functional pulmonary venous stenosis. Since the interventional ablation strategies for atrial fibrillation target the veno-atrial junctions of the pulmonary veins, plenty of cases with haemoptysis due to post-procedural pulmonary vein stenosis were reported. Due to thermal injury at the veno-atrial junctions [68, 69], pulmonary venous stenosis has become a well-recognised complication of these ablation procedures, with sometimes even fatal peri-procedural haemoptoe [41]. Because of this potentially life-threatening complication, ablation technology has changed from the intra-ostial radiofrequency ablation technique to the antral cryoballoon isolation technique during the past decade. With the radiofrequency procedures, pulmonary vein stenosis rates of up to 30–40% in scheduled surveillances were reported. With the cryoballoon technique, the official pulmonary vein stenosis rate is about 1–3% [70]. In the USA alone, about 40 000–50 000 atrial fibrillation ablation procedures are performed each year. This review focuses on haemoptysis due to the obstruction of the central mediastinal pulmonary veins.
Anatomy of the mediastinal pulmonary veins
The embryological formation of the pulmonary veins is quite complex and there is a remarkable diversity of pulmonary vein connections into the left atrium [71–74]. The classical description of two left-sided (left superior (LSPV) and left inferior (LIPV)) and two right-sided (right superior (RSPV) and right inferior (RIPV)) separately ending pulmonary veins accounts for only ∼70% of the normal population. Approximately 28% of people have three to five ostia at the right side and 14% of the population have a single ostium at the left side. At the right side an accessory right pulmonary vein is common. The right-sided middle lobe vein (RMLV) may drain via the superior pulmonary vein, the inferior pulmonary vein or separately. The variant of RMLV drainage via the RIPV may result in life-threatening post-operative haemoptysis and respiratory distress in case of right lower lobectomy. Therefore, careful pre-operative assessment of the pulmonary venous return is crucial in patients undergoing pulmonary surgery [24, 75, 76]. In addition to congenital stenosis of pulmonary veins with normal connection, the varying forms of partial anomalous venous return may be combined with a pulmonary venous stenosis.
Pathophysiology
To understand the pathophysiology of haemoptysis in cases of pulmonary venous stenosis, it is important to know that both the pulmonary and the bronchial circulation drain via the pulmonary veins into the left atrium. Thus, in pulmonary venous stenosis the drainage systems of both lung circulations are blocked. Typical consequences include distended pleural-hilar bronchial veins, alveolar haemorrhage, a friable endobronchial mucosa, a reduced lymphatic drainage, interstitial pulmonary oedema, enlarged hilar lymph nodes, enlarged lymph vessels and sometimes pleural effusions [77–82]. To keep the lung an optimal gas-exchanging system, the pulmonary arterial blood flow is also affected, with redistribution of the pulmonary arterial blood flow towards regions with lower vascular resistance [83–87]. In severe stenosis even a reversal flow in the pulmonary arteries with development of pulmonary venous hypertension, pulmonary arterial remodelling and a decreased arborisation of the pulmonary arterial tree may develop [88, 89] (table 2).
Haemoptysis and other clinical symptoms and signs caused by pulmonary venous stenosis
Depending on the acuity of the pulmonary venous obstruction and the development of venous collaterals the venous parenchymal and mucosal bleeding may vary from even asymptomatic courses with occult alveolar haemorrhage to acute fatal haemoptoe. In cases of haemoptysis due to pulmonary venous stenosis the expectorated blood is deoxygenated and, therefore, usually darker compared to haemoptysis with offspring of the systemic bronchial arteries. As in other cases with massive haemoptysis, clinical management includes early endotracheal intubation with large-bore tubes in an intensive care unit setting, early bronchoscopy for localisation of the bleeding side and early endobronchial therapy to protect the nonbleeding side. The endobronchial changes with dilatation of the dense submucosal venous plexus can often be seen in bronchoscopy and the alveolar haemorrhage will result in a bloody bronchoalveolar lavage or, if occult, in an increased number of siderophages in the cell differentiation [82]. Due to the dense network between the pulmonary and bronchial circulation, extensive collaterals between both circulations may develop, with the possible occurrence of secondary bronchial and pulmonary venous varices in the long run. Misinterpretation of these varices and collaterals as pulmonary arteriovenous malformations has been reported. Bronchial artery embolisation may be deleterious in cases of hindered pulmonary venous drainage.
Besides haemoptysis, the typical clinical symptoms of pulmonary venous stenosis are tussive irritation, exertional dyspnoea, recurrent pulmonary infections and signs of pulmonary venous hypertension. Mimicking of primarily pulmonary diseases, such as recurrent pneumonia, interstitial pneumonitis, lung cancer and pulmonary embolism, has been reported in animal studies, as in many case reports [19, 90–93]. Mild interstitial fibrosis is thought to originate from the organisation of haemorrhagic oedema in the alveolar walls. The resulting clinical signs of pulmonary venous stenosis are highly unspecific and may, therefore, be misleading. Unilateral pleural effusions, haemoptysis in combination with enlarged hilar lymph nodes and pulmonary infiltrates in combination with positive blood cultures are only a few examples of the consequences of pulmonary venous stenosis. Figure 1 demonstrates different clinical signs caused by pulmonary venous stenosis in a patient with life-threatening haemoptysis. Haemoptysis in combination with the history of mediastinal masses, granulomatous diseases, interventional procedures for atrial fibrillation therapy, lobectomy or right-sided pneumonectomy, pulmonary hypertension, recurrent respiratory infections or after lung transplantation should prompt the suspicion of pulmonary venous stenosis (table 1). Yet the diagnostic workup of pulmonary venous stenosis is elusive, which is why the different diagnostic procedures to detect pulmonary venous stenosis are discussed here.
a) Chest radiograph, b) multislice computed tomography angiography, c) three-dimensional magnetic resonance angiography, d) perfusion scintigraphy and e) iron staining of a lung biopsy of a patient with severe haemoptysis due to occlusion of the right-sided pulmonary veins. Note the right-sided interstitial infiltrates, the hilar lymphadenopathy, the pleural effusion, the perfusion shift to the left lung and the siderophages in the specimen (a–e, respectively).
Diagnostic imaging
Accurate imaging of the anatomical and functional properties of the pulmonary veins is challenging. For example, a pathologic perfusion scan in combination with haemoptysis may be misinterpreted as acute pulmonary embolism [94, 95]. Various imaging techniques have been investigated to explore the mediastinal pulmonary veins, such as transoesophageal echocardiography (TOE), ventilation/perfusion scan, contrast-enhanced multisclice CT, magnetic resonance angiography and direct pulmonary venography. Wood et al. [96] compared CT, TOE and direct pulmonary venography to assess the number of detectable pulmonary veins and their diameters at the veno-atrial junctions in 24 patients. In contrast to CT, which detected 98 pulmonary veins, TOE detected only 80 and direct venography detected only 71 pulmonary veins. Thus, direct angiography missed 27% and TOE missed 18% of the pulmonary veins. With TOE in particular, the inferior and the right middle lobe veins will be missed. Depending on which imaging technique is used, the ostial diameters vary significantly from 1.6±0.3 cm using direct venography to 1.1±0.25 cm using TOE. Therefore, neither TOE nor direct pulmonary venography can be recommended for the exclusion of pulmonary venous stenosis. Depending on which imaging technique is used, the ostial diameters vary significantly from 1.6±0.3 cm using direct venography to1.1±0.25 cm using TOE.
A brief discussion of the pros and cons of these imaging techniques referring to pulmonary venous stenosis is shown in table 3 and discussed below.
Ventilation/perfusion scan
The ventilation/perfusion scan usually is performed for the detection of pulmonary embolism, but is also reported to serve as an effective screening tool for the detection of haemodynamically relevant pulmonary venous stenosis [84–87]. To maintain the lung, an optimal gas exchanging system, backward transmission of elevated venous pressures will immediately induce a local increase of vascular resistance with a subsequent shift of the pulmonary arterial blood flow towards regions with lower resistance. This shift explains the perfusion deficits in lung perfusion scans of patients with haemodynamically relevant pulmonary venous stenosis. Lepădat et al. [87] demonstrated an immediate shift of the pulmonary perfusion after pulmonary arterial as well as pulmonary venous ligation. Nanthakumar et al. [84] demonstrated pathological perfusion scans with a resting pressure gradient >5 mmHg between the pulmonary vein and the left atrium, or with a ≥80% luminal stenosis. To minimise false-positive pulmonary embolism results, lung perfusion scans are best performed using the technique of single photon emission computed tomography (SPECT) and should be interpreted in conjunction with the corresponding ventilation scan. However, in case of alveolar haemorrhage and endobronchial thrombus formation, pulmonary venous stenosis may also result in a combined ventilation/perfusion deficit. Perfusion deficits may be missed if the stenosis is <50%.
Magnetic resonance imaging
Magnetic resonance imaging (MRI) offers multiple advantages such as noninvasiveness, freedom from radiation exposure and visualisation of the circumadjacent mediastinal structures. Furthermore, it provides information about blood flow and left and right ventricular function. For the assessment of pulmonary venous stenosis different MRI techniques are available, such as noncontrast white-blood imaging, three-dimensional (3D) steady-state free precession magnetic resonance angiography, four-dimensional flow MRI, phase-contrast velocity mapping and contrast-enhanced gadolinium magnetic resonance angiography [88, 89, 97–101]. Magnetic resonance-perfusion imaging has 95% sensitivity and 100% specificity to detect perfusion deficits in cases of haemodynamically relevant pulmonary venous stenosis, compared to the scintigraphic SPECT technique [97]. With phase-contrast MRI, the pulmonary venous flow and mean flow velocities in the mediastinal veins can be quantified. Signs of venous obstruction include flow acceleration downstream from the stenotic lesion and a loss of the normal phasic changes. Normal systolic and diastolic peaks of the pulmonary venous blood flow were reported at 51±16 cm·s−1 and 47±11 cm·s−1. There is a significant variability in pulmonary vein diameter and cross-sectional area over the cardiac cycle, with greatest diameters in ventricular diastole and a mean difference in diameter of ∼15% and 27% for the cross-sectional area. For optimal imaging of the pulmonary veins, the authors suggest a gadolinium-enhanced, bolus-tracking technique with the region of interest in the left atrium to match the imaging sequences with the maximum contrast media peak into the left atrium.
Computed tomography
The advantage of multislice CT-venography compared to MRI is the simultaneous visualisation of the pulmonary parenchyma and that 3D-reconstructions of the pulmonary veins can easily be performed. CT scanning provides additional information about the functional consequences of pulmonary venous stenosis, because the morphological parenchymal alterations, such as thickening of the septal walls, mosaic pattern due to inhomogeneous aeration and perfusion, may not be detected with MRI. As with MRI, the scanning delay time should be defined by a bolus-tracking technique focusing the bolus arrival into the left atrium for optimal visualisation of the veno-atrial junctions. Mean pulmonary vein diameters were reported as follows: RSPV 13–15 mm, LSPV 16–17 mm, RIPV 16–17 mm, LIPV 14–15 mm and middle lobe vein 8.2–8.9 mm. Up-to-date (64-row or 128-row) multislice-detector CT scans are able to visualise the veno-atrial junctions with excellent spatial resolution [102–107].
Echocardiography
The transthoracic echo window usually is unsatisfactory for the evaluation of the pulmonary venous flow in adults. TOE, with the use of colour- and Doppler-mode at a Nyquist limit of ∼40 cm·s−1, offers a much better ultrasound window. For the pulsed-wave Doppler analysis the standard position of the sample volume is about 1–2 cm upstream from the pulmonary vein orifices, where normally a triphasic flow pattern can be delineated. This flow profile is hooked to the left atrial pressure with mean velocities of ∼0.5 m·s−1 in systole and 0.4 m·s−1 in diastole [108]. Pulmonary venous stenosis is suspected if peak flow velocities exceeds 1.0 m·s−1 and/or if pulmonary vein diameter is <5 mm. The inferior pulmonary veins and accessory right pulmonary veins are often difficult to view because Doppler imaging is highly angle dependent. Visualisation of the pulmonary veins with TOE is much poorer compared to the newer generations of multislice imaging techniques [96].
Direct venography
Direct venography is the most invasive and time-consuming method and is restricted to a planar luminography. Furthermore, a trans-septal puncture is required. However, in the case of a near-total pulmonary vein occlusion, a flow-through pulmonary angiogram sometimes demonstrates a residual pulmonary vein lumen, which then facilitates a retrograde recanalisation of the pulmonary vein [10]. Except in the case of a near-total occlusion, the multislice techniques have a much higher diagnostic accuracy, which is why the direct pulmonary venous angiography is performed in the context of therapeutic procedures rather than for diagnostic reasons.
Therapeutic options to restore pulmonary venous drainage in the case of central pulmonary venous stenosis
If pulmonary venous stenosis is diagnosed, catheter-guided and surgical procedures are established for therapy, depending on the aetiology of the stenosis. Pulmonary venous stenosis is usually a progressive disease resulting in a kind of functional lobectomy due to the perfusion shift, which is why symptomatic higher stenosis should be treated. In the case of pulmonary venous stenosis at the veno-atrial junctions due to interventional therapy of atrial fibrillation, a percutaneous catheter-guided dilatation is recommended. After puncture of the vena femoralis using the Seldinger technique and subsequent atrial transseptal puncture, a guiding catheter is placed in the left atrium under fluoroscopy. Then, the interventionalist places a guidewire into the stenosed pulmonary vein and dilates the venous stenosis with a balloon catheter. Surgical procedures are commonly used in cases of congenital stenosis of the veno-atrial junctions either with conventional or with sutureless venoplasty or a pericardial patchplasty [109]. For aquired pulmonary venous stenosis due to mediastinal masses, the therapeutic decision is usually made subject to individual aspects. Both surgical techniques, as with catheter-guided techniques, have the problem of an extremely high proportion (up to 50%) of restenosis. Catheter-guided stent implantation with bare-metal stents or drug-eluting stents may be an alternative if restenosis develops [110]. If massive bleeding persists, an urgent surgical lobectomy or pneumonectomy can be life saving. Anticoagulation is mandatory after successful recanalisation of pulmonary venous stenosis; however, in the acute setting with haemoptysis anticoagulation may be detrimental. table 4 suggests a diagnostic algorithm for patients with unexplained haemoptysis, to exclude pulmonary venous stenosis of the mediastinal pulmonary veins and the veno-atrial junctions, and it describes the therapeutic options to restore the pulmonary venous drainage.
Conclusions
Pulmonary venous stenosis should be considered in the differential diagnosis of patients with haemoptysis. In particular, patients with a history of mediastinal masses, interventional therapy for atrial fibrillation or right lower lobectomy are predisposed to pulmonary venous obstruction. TOE, ventilation/perfusion scans and direct venoography may detect pulmonary venous stenosis, but cannot be recommended for the exclusion of pulmonary venous stenosis. Gadolinium-enhanced magnetic resonance venography and contrast-enhanced multislice CT with 3D reconstruction of the veno-atrial junctions are the most appropriate methods for the detection of pulmonary venous stenosis. For optimal visualisation of the veno-atrial junctions, the correct scanning delay time is important.
Footnotes
This article has supplementary material available from err.ersjournals.com
Provenance: Submitted article, peer reviewed.
Conflict of interest: None declared.
- Received May 19, 2013.
- Accepted September 6, 2013.
- ©ERS 2014
ERR articles are open access and distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0.
References
