Pulmonary artery aneurysms (PAAs) are well-known causes of mortality and morbidity in Behçet disease (BD).13,22–24,31,50,51 However, pulmonary artery involvement is not limited to PAA. The other main type of pulmonary artery involvement in patients with BD is pulmonary artery thrombus (PAT).50,51 In recent years, we have repeatedly observed patients with pulmonary artery involvement presenting with additional lung lesions such as nodules, consolidations, and cavitations. Diagnosis and management of these lesions can be difficult, and they are frequently misinterpreted as being due to infection, especially in the setting of immunosuppressive use as is usually the case in BD patients with pulmonary artery involvement. Such patients have been reported in only a few case reports.20,21 It has been reported that the prognosis of BD patients with pulmonary artery involvement presenting as isolated PAT was better than the prognosis of those with PAA, while vascular involvement in BD could be divided into micro- and macrovascular forms.51
For the current study we surveyed a group of BD patients with pulmonary artery involvement diagnosed and followed in our dedicated center between 2000 and 2007. We studied the clinical characteristics, vascular and nonvascular radiologic lung lesions, echocardiographic findings, pathologic findings if available, clinical course, and prognosis, along with treatment modalities. We specifically focused on similarities and differences between the PAA and PAT type of involvement. We also formally analyzed, we believe for the first time, the relationship between parenchymal lesions and pulmonary artery involvement. (For more information about BD, see the articles in this same issue on arterial lesions43a and cardiac involvement18a in BD.)
PATIENTS AND METHODS
We reviewed records of about 2500 patients with BD who were registered at the multidisciplinary clinic at Cerrahpasa Medical Faculty between January 2000 and December 2007. From this group we identified 47 (41 male, 6 female) patients who were diagnosed with pulmonary artery involvement during this period. Pulmonary artery involvement was defined as having either PAA, PAT, or both in combination strictly diagnosed by thoracic computed tomography (CT) or magnetic resonance imaging (MRI).
Thirty-six of these patients had been referred to or presented at our outpatient clinic for the first time because of pulmonary symptoms: 27 had PAA with or without PAT, and 9 had isolated PAT. The remaining 11 (23%) patients were already being followed for other symptoms of BD. The end of follow-up was set as June 2010.
We collected the following information: demographics, clinical history (including signs and symptoms at diagnosis and during the course of the disease), imaging findings (thoracic CT scan and MRI, vascular Doppler, and positron emission tomography [PET] scan), laboratory data, echocardiographic examinations, and detailed medical and surgical treatments. Only the initial chest radiographs at the time of presentation were also re-evaluated.
By the end of the survey (June 2010), 12 (11 male, 1 female) patients had died. Of the remaining 35 patients, 31 (27 male, 4 female) were still attending the outpatient clinic regularly and had been seen within the preceding year. In only 4 (3 male, 1 female) patients, we obtained outcome information from the primary physicians.
In addition to the data from the structured chart survey, during their last clinical visit the patients were questioned again about the occurrence of clinical signs and symptoms (hemoptysis, with massive hemoptysis being defined as hemoptysis >500 cc; cough; pleuritic chest pain; shortness of breath with exercise; and fever) within the previous year.
All available thoracic CT scans that had been taken from the onset of pulmonary symptoms to the end of follow-up were systemically re-evaluated by 2 experienced radiologists (CA and HO), who were aware of the purposes of the study but did not know the clinical history of the examined patients. In case of disagreement, the decision of the senior radiologist (CA) was followed.
We used a predefined checklist to evaluate the changes in pulmonary arteries, pulmonary parenchyma, intrathoracic veins, mediastinum and lymph nodes, pleura, heart, and pericardium. (See the checklist, the Thoracic CT Evaluation Form, in the Appendix, available online as Supplemental Digital Content, http://links.lww.com/MD/A8) Arterial changes included aneurysms, thrombosis, stenosis, occlusions, cutoffs, and oligemia (decreased vascularization) located throughout the pulmonary arterial tree. The following lesions were assessed in the lung parenchyma: nodular infiltrations, consolidations, cavities, and ground-glass opacities. Lymphadenopathy, pleural effusion or thickening, intracardiac filling defects, and pericardial effusions or thickenings were also noted. The anatomic location, shape, and size of each lesion—parenchymal, mediastinal, or vascular—were specified.
The initial chest radiographs were available in 39 of 47 (83%) patients. These were re-evaluated for the presence of hilar opacities, nodular lesions, and pleural effusions at the time of presentation to the clinic.
Continuous variables with a more or less normal distribution were expressed as means ± standard deviations (SDs), and others as medians and interquartile ranges (IQRs). Categorical variables were compared using the chi-squared test or Fisher exact tests, as appropriate. Comparisons of continuous variables were made using the Student t test. Continuous variables with non-normal distributions were compared using the Mann-Whitney U and Kruskal-Wallis tests with Bonferroni correction for multiple pairwise comparisons. Survival data were assessed by Kaplan-Meier analysis. All tests were performed using version 13.0 of SPSS software for Windows (SPSS Inc., Chicago, IL).
Demographic and General Clinical Characteristics
Demographic and clinical characteristics are given in Table 1. Patients were mostly male (41/47; 87%). Mean age was 29 ± 8 years, while mean disease duration until the onset of pulmonary artery involvement was 3.6 ± 4.8 years (median, 2 yr). Mean follow-up time for patients who survived (n = 35) was 6.6 ± 2.0 years; for those who died (n = 12) it was 1.5 ± 2.3 years.
Four patients did not fulfill the International Study Group (ISG) criteria28 at the time of diagnosis of pulmonary artery involvement. One was a 12-year-old boy who had recurrent oral ulcers, erythema nodosum, hepatic vein thrombosis, and negative pathergy test. He died after 1 year of follow-up. The second patient was a 26-year-old man who only had deep vein thrombosis in the legs and an intracardiac thrombus. He started to have oral and genital ulcers 1 year after the onset of pulmonary artery disease, and later developed thrombosis in the superior vena cava and hepatic veins. The third was a 30-year-old woman who had had only recurrent oral ulcers, superficial thrombophlebitis, and arthralgia for 2 years when she developed PAA. She still did not fulfill the ISG criteria after 5 years of follow-up. The fourth patient was a 27-year-old man who had only recurrent oral ulcers and extensive venous thrombosis including lower extremity veins, vena cava inferior, and right atrium when he developed PAT. He also had a brother with BD who had eye disease. After 3 years of follow-up he still had not fulfilled the ISG criteria.
Venous thrombosis was the most frequent accompanying lesion (36/47) (see Table 1). Only 2 male patients had arterial involvement outside the lungs: 1 had diffuse arterial wall thickening in the brachiocephalic and bilateral carotid arteries similar to Takayasu arteritis, and the other had a cerebellar artery aneurysm.
For the most part the demographic and clinical characteristics of the patients did not differ significantly between the 2 types of pulmonary artery involvement (see Table 1). However, eye disease (p = 0.002) and perhaps dural sinus thrombosis (p = 0.059) were less frequent among the patients with PAA compared to patients with isolated PAT. At the time of pulmonary artery involvement diagnosis, 43 (91%) patients also had active BD affecting other parts of the body: 11 of 43 (23%) patients had active uveitis, 13 (28%) had acute venous disease—not to be confused with acute and/or chronic peripheral venous disease present in 36 of 47 (77%) in the whole group as noted above—and 40 (85%) had active skin-mucosa lesions.
Of the 47 patients, 3 had been using azathioprine (AZA), 2 cyclosporine, 8 colchicine, and 2 corticosteroids alone, when pulmonary artery involvement developed. Work-up for infection before the diagnosis of pulmonary artery involvement had been carried out in 28 (60%) patients. This included sputum Gram and acid-fast stains (n = 25) in addition to blood (n = 28) and bronchoalveolar lavage fluid (n = 4) cultures for bacterial, mycobacterial, and fungal infection.
Hemoptysis was the most common presenting symptom of pulmonary artery involvement followed by cough, fever, dyspnea, and pleuritic chest pain (Table 2). Hemoptysis was massive (>500 cc) in 17 of 47 (39%) patients. As a presenting symptom, hemoptysis was significantly more common in patients with PAA (31/34, 91%) than in patients with isolated PAT (6/13, 46%) (p = 0.001) The same was true for massive hemoptysis (p = 0.008). Apart from hemoptysis, presenting symptoms were not significantly different between the 2 types of pulmonary artery involvement. Fever was present in 22 patients, and in 17 (77%) was an early sign, preceding other symptoms.
Follow-Up and Outcome
At the end of the study, 16 of the 35 (46%) patients who remained alive were symptom free. The remaining patients continued to have at least 1 symptom (dyspnea on exertion, n = 13 [37%]; small bouts of hemoptysis, n = 8 [22%]; cough, n = 5 [16%]; and chest pain, n = 1). Hemoptysis if still present at the time of final evaluation was only in small amounts; either <100 cc or in the form of blood-tinged sputum. Unlike other symptoms which decreased in frequency with time, exercise-induced dyspnea increased in frequency at final evaluation (23% vs. 38%, p = 0.192). The level of dyspnea that patients described was New York Heart Association class II.
Pulmonary Arterial Findings
Thirty-four (72%) patients had PAA (Figure 1); 8 of these had concomitant PAT (Figure 2). Thirteen (28%) patients presented with isolated PAT (Figure 3). Three of these 13 patients developed PAA along the thrombosed arteries after 6, 10, and 24 months of follow-up. None of the 8 patients who had both PAA and PAT at presentation developed additional PAA at other sites.
The involvement was bilateral in 27 (57%) patients. The numerical details of the arteries involved as well as the type of involvement are tabulated in Table 3. The descending branches of the pulmonary arteries were the most common place of involvement; this was true for both PAA and PAT. A mural thrombus was seen in the majority of the aneurysms (83/97; 85%).
The mean diameter of aneurysms was 2.3 ± 1.1 cm. PAA were saccular, fusiform, or tubular. In 2 patients, PAA had caused right middle lobe atelectasis by obliterating the nearby bronchus. Air bubbles suggesting bronchial connection could be seen on the periphery of large aneurysms.
The initial chest radiographs at the time of onset of the pulmonary symptoms were available for 39 patients (28 of 34 patients with PAA and 11 of 13 patients with isolated PAT). Significantly more initial chest radiographs were normal in patients with isolated PAT (6/11) than in patients with PAA (1/28) (p < 0.001). In the remaining patients with PAA, chest radiographs showed hilar opacities (n = 20), peripheral nodules or consolidations (n = 8), and pleural effusion (n = 11). On the other hand, among the remaining patients with isolated PAT, there were pleural effusions along with consolidations in 4 patients and a large cavity in another.
Of the 34 patients with PAA with or without PAT, 9 (26%) died a median of 4 months (IQR, 2-17 mo) after diagnosis (Figure 4). In 23 (68%) who survived, aneurysms regressed in 15 days at the earliest, and disappeared after a mean of 10 ± 7 months, in 2 months at the earliest. In 2 patients, all aneurysms disappeared except for a single aneurysm in each patient that diminished in size but persisted (1.5 × 1.1 mm and 1 × 0.8 mm, respectively). Aneurysms recurred in 4 other patients 8, 21, 22, and 29 months after disease onset, respectively, but again disappeared completely. Of these 4, 2 patients relapsed twice. Apart from these, 1 patient who presented with both PAA and PAT (Patient 10, Table 5) continued to have progressive thrombosis until his death (described below in the section on patients with parenchymal relapses).
Of the 13 patients with isolated PAT at initial presentation, 10 continued to have isolated PAT during the follow-up, while 3 patients developed aneurysms after 6, 10, and 24 months of follow-up, respectively. One died and 2 survived with complete disappearance of the aneurysms. Of the remaining 10 patients with isolated PAT who did not develop PAA, 2 died and 8 survived, among whom only 1 suffered a thrombotic relapse.
In total, pulmonary artery involvement relapses were seen in 9 (19%) patients (PAA: 5, PAT: 4).
By the end of follow-up, of the 35 patients still alive, arterial lesions (PAA or PAT) healed with no arterial sequelae in 14 (40%) patients (PAA: n = 12, isolated PAT: n = 2). Twenty-one patients (PAA: n = 13, isolated PAT: n = 8) (60%) had arterial stenosis or occlusions, mainly in the lower lobe arteries. Among these same 21 patients, 2 patients had a single and persistent aneurysm (as described above).
The frequency of parenchymal lesions did not differ between patients presenting with PAA or isolated PAT (Table 4).
Nodular lesions (Figure 5) were more likely to be found during the acute phase of pulmonary artery involvement; such lesions were present in 40 (85%) patients at presentation with pulmonary artery involvement. Thirty-three (83%) of these patients were not receiving corticosteroids or immunosuppressive treatment when these lesions were first detected. In 7 patients (7/40, 15%) these lesions were found in the thoracic CT scans, when pulmonary symptoms such as fever, coughing, and chest pain had begun and CT scans were still negative for pulmonary artery involvement.
As seen in Table 4, the lesions were mostly bilateral (23/34, 68%), and were predominantly found in the inferior lobes (73%). Nodular infiltrations or consolidations were usually distributed in the peripheral area and were described often as subpleural and seldom as pleural based.
Open lung biopsies of peripheral parenchymal nodules were available for 5 patients. Pathologic evaluation of the biopsies revealed both infarction and necrosis (n = 3) (Figure 6A and B), organizing pneumonia (n = 1) (Figure 6C), and necrotizing granulomatous inflammation (n = 1) (Figure 6D). All biopsies were performed after immunosuppressive treatment had been initiated.
Follow-Up and Outcome
Two different forms of lesions were observed. Most of the lesions were quick resolving, usually in 2–3 weeks, only to appear frequently at other sites (wandering), which similarly lasted for short periods. On the other hand, there were also some slow healing, mostly pleura-based lesions persisting for longer periods, such as 6–15 months. Less often, bulky lesions similar to consolidations were observed. Either type when finally resolved was replaced by fibrotic scars or disappeared completely. During the follow-up, 11% (18/171) of the nodules transformed into cavitary lesions.
The number of patients with nodular infiltrations decreased significantly when initial and final CT scans were compared (40/47 vs. 6/35, respectively, p < 0.001).
At Presentation and Follow-Up
Cavitary lesions were found in 22 of 47 (47%) patients (Figure 7). They were already present at presentation in 6 of 47 patients (13%), while in the remaining 16 (34%) patients they appeared after a mean of 14 ± 8 months. Four of the 6 patients with cavities at initial presentation were not receiving corticosteroids or immunosuppressive treatment.
A total of 18 cavities in 10 patients developed from excavation in large nodular infiltrations or consolidations. They usually had thin walls and well-defined borders. Similar to what was observed in pulmonary artery involvement and the nodular infiltrations, they were mainly found in the inferior lobes (see Table 4). Nine patients (9/22, 41%) had cavities with air fluid levels. While no microbial agent could be demonstrated, patients usually received empirical antibiotics in addition to immunosuppressive agents.
At the final evaluation, cavities had disappeared in 12 patients. In 4 patients they had regressed in size but still persisted. Two (2/43) giant cavities in 1 patient were surgically repaired, and 2 (2/43) cavities (2 patients) had caused pneumothorax. The outcome of 5 cavities in 3 patients who had died was unknown.
Other Lesions (Ground-Glass Opacities, Pleural and Pericardial Effusion or Thickening, Lymphadenopathy)
Ground-glass opacities were observed in 21 of 47 (45%) patients. In 9 patients they were attributed to intraparenchymal hemorrhage (Figure 8). A bilateral large pleural effusion, a transudate, was observed in only 1 patient. This was associated with acute vena cava superior thrombosis, which had occurred concomitantly with PAA. Minimal pleural effusions were observed in 20 (43%) patients, and these were usually accompanied by nodular lesions. The character of the fluid could not be examined because its quantity was small. Minimal pericardial effusion or thickening was detected in 10 (21%) patients.
Mediastinal lymphadenopathy was documented in 10 of 47 (21%) patients, mainly at the initial CT scans. These were mainly paratracheal, subcarinal, paraesophageal, and hilar lymph nodes, with a diameter of 1.5 cm and less. Calcified lymph nodes were found in 1 patient, who had a previous history of lung tuberculosis.
Follow-Up and Outcome
Similar to nodular infiltrations, these lesions attenuated or disappeared during the follow-up.
Intracardiac Filling Defects
Intracardiac filling defects compatible with thrombi were detected in 13 (28%) patients (Figure 9), located in the right ventricle (n = 11) and right atrium (n = 2). Two patients had 2 lesions; the rest were solitary. The maximum diameter of the defects ranged between 10 and 50 mm. Echocardiography done simultaneously with the CT scans detected intracardiac thrombi in 8 (62%) of these patients.
Follow-Up and Outcome
Final follow-up CT scans were available for 11 of 13 patients and showed that these defects were much smaller and decreased in intensity in 9 patients after a median follow-up of 2 years. The filling defects calcified in time in the remaining 2.
Patients With Relapses and Parenchymal Findings Suggesting Infection
A subset of patients (n = 7) relapsed under the standard treatment protocol (described below) with symptoms like fever, cough, increase in acute-phase reactants, and parenchymal lesions seen in CT scans such as nodular infiltrations and cavitary lesions (Figure 10). As briefly mentioned in the Introduction, the complex and difficult clinical course, especially the difficulty we had in ruling out lung infection in these patients, was the main impetus for this survey.
At the time of relapse, the activity of the arterial disease showed wide variance: PAA or PAT had disappeared in 2 patients, regressed in 1, persisted in 2, and relapsed in 2. All 7 patients were receiving monthly pulses of cyclophosphamide (CYP) in addition to oral prednisolone which was being tapered when the relapses occurred. It was difficult to differentiate the clinical picture of these patients from that seen in patients with an opportunistic lung infection, as these patients were under immunosuppression. All 7 patients were hospitalized and underwent a detailed investigation to exclude infection: sputum and bronchial lavage fluid were collected for Gram and acid-fast stain along with bacterial, mycobacterial, and fungal cultures. Blood cultures were done for aerobic and anaerobic bacteria, Mycobacterium tuberculosis, and fungal infection. Blood PCR tests were done for the detection of M. tuberculosis complex and cytomegalovirus infection. Infection work-up also included tuberculin test. Despite this intense work-up, no infectious etiology could be shown, and patients were unresponsive to empirical treatment with wide-spectrum antibiotics. However, when high doses of prednisolone were administered and CYP pulses were restarted, the attacks subsided in all but 1 patient, who had presented with PAA and PAT. Aneurysms disappeared but he continued to develop thrombotic occlusions in the pulmonary arteries despite intensive immunosuppressives and prednisolone. He had persistent fever, severe dyspnea, and multiorgan failure. The patient ultimately died, probably of sepsis (Patient 10, Table 5). Pneumothorax developed subsequently in another patient, who was treated with chest tube insertion. Another patient who had enlarging cavities required an emergency decortication and correction of the bronchopleural fistula. He, too, improved and was still alive at the end of follow-up.
At diagnosis of pulmonary artery involvement, median erythrocyte sedimentation rate (ESR) was 67 mm/1 h (IQR, 48–82 mm/1 h) and C-reactive protein (CRP) level (normal limits <5 mg/L) was 62 mg/L (IQR, 24-83 mg/L). ESR and CRP levels did not differ between those with isolated PAT and those with PAA (data not shown).
Follow-Up and Outcome
Among those who survived, ESR and CRP were significantly decreased at final assessment (ESR: median, 35 mm/1 h [IQR, 4–19 mm/1 h]; CRP: median, 8.3 mg/L [IQR, 1–4.9 mg/L]; for both ESR and CRP, p < 0.001).
At initial presentation echocardiography was obtained in 36 of 47 (76%) (32 male, 4 female) patients. Systolic pulmonary artery pressure (sPAP) was measured above the cut-off level of 30 mm Hg in 23 (65%) of the patients. The mean value was 39 ± 18 mm Hg (range, 30-65 mm Hg). Among 36 patients with echocardiograms, intracardiac thrombi were observed in 12 (33%) (10 male, 2 female) patients, localized in the right ventricle (n = 9) or the right atrium (n = 3). The frequency of intracardiac thrombi did not differ significantly between those with PAA (8/25, 32%) and those with isolated PAT (4/11, 36%).
Follow-Up and Outcome
Echocardiographic studies were repeated in 22 patients (18 male, 4 female) after a median of 3 years (IQR, 2–4 yr). sPAP did not appreciably change in 6 of 11 patients with an initially high value (range, 35–45 mm Hg), but it returned to normal levels in the remaining 5 patients. sPAP remained within the normal range in 11 other patients during the follow-up. Of the 8 patients with intracardiac thrombi, the lesions resolved in 5 and in the remaining patients the intracardiac echogenic lesions were interpreted this time as fibrotic lesions rather than true thrombi.
Ventilation perfusion (V/Q) lung scans were available in 24 (51%) patients. Bilateral multiple V/Q mismatched areas suggesting pulmonary embolism with high probability52 were found in 19 (79%) patients, intermediate in 3 (13%), and low in 2 (8%). The lung scan of a patient who had only mild dyspnea showed ventilation but no perfusion at all in the right lung. A CT of the lung in the same patient showed total occlusion of the right main artery, dilated bronchial arteries on the right side, relatively shrunken right lung, and probable compensatory enlargement of the pulmonary arteries on the left side.
V/Q scans were repeated in 6 patients because of continuing mild dyspnea after 2–6 years of follow-up. Four patients had, as the rest of the group also had, peripheral major vein thrombosis. The repeat scans continued to show the same abnormalities as the initial scans.
PET scans were performed in 3 patients and showed significant fluorodeoxyglucose (FDG) uptake in lung segments matching the aneurysms, parenchymal consolidations, and mediastinal lymph nodes detected on CT scans.
Our usual treatment protocol22,23 has been CYP given as 1 g monthly pulses for at least 1 year always combined with prednisolone. Prednisolone has been administered as 3, 1 g IV pulses in addition to oral 1 mg/kg per day dose. It was tapered by 10 mg/mo after the first month and usually continued for at least 1–2 years 5–10 mg/d.
A total of 41 patients (PAA: 32, isolated PAT: 9) received CYP as first-line treatment after the diagnosis of pulmonary artery involvement. In the remaining patients, AZA (n = 4) with anticoagulants (n = 2) and prednisolone only (n = 2) were used initially. The 2 patients in whom anticoagulants had been used had isolated PAT. One received anticoagulation before referral to our center, and the other had elevated antiphospholipid antibodies. Of the 3 patients with isolated PAT who developed PAA during the follow-up, 2 were receiving AZA while 1 was under monthly CYP. All 3 received CYP and high-dose corticosteroids after being diagnosed with PAA. Thus, 43 patients overall were treated with CYP. The median duration of CYP use was 14 months (IQR, 10–24 mo) among those who remained alive (32/35 patients) and 3 months (IQR, 2–5 mo) among those who died (11/12 patients). One patient who had recurrent hemoptysis while treated with CYP responded well to infliximab. After induction treatment with CYP, AZA was the most commonly used maintenance drug among these patients (30/35, 89%), followed by interferon (3/35) and colchicine (2/35). However 9 patients had to stop AZA due to side effects like cytopenia, rash or gastric pain (n = 6) or relapse (n = 3). The median duration for AZA use was 2.6 years (IQR, 1–5 yr).
By the end of the study, while 26 were still using immunosuppressant agents or corticosteroids (AZA: n = 20, interferon: n =1, mycophenolate mofetil: n = 1, infliximab: n = 1, colchicine: n = 1, corticosteroids alone: n = 2), 9 patients did not use any treatment.
We observed that pulmonary artery involvement recurred in 9 patients in total, while receiving CYP (n = 3), AZA (n = 5) and no treatment (n = 1).
Other Treatment Procedures (Embolization and Surgery)
Two patients (1 male, 1 female) underwent embolization and another male patient had a lobectomy. Apart from these, 2 others (both male) had several surgical operations because of complicated cavities and refractory pleural effusions.
An embolization with N-butyl cyanoacrylate (NBCA) was performed in a female patient who had bilateral giant PAAs (7 × 4 cm and 5 × 4 cm) that had not responded to standard immunosuppressive treatment. The procedure was unsuccessful, in that the embolization material did not occlude the aneurysms and an air cavity developed following the procedure. The patient died subsequently (Patient 9, Table 5).
Coil embolization was done in 2 aneurysms (5 × 4 cm and 2 × 2 cm in size, respectively) in another patient. The procedure was unsuccessful in the larger aneurysm. The aneurysm sac was transformed into an air cavity, and a septic empyema developed following the procedure. The coil material was surgically removed in addition to decortication and bronchopleural fistula repair about 4 months after the embolization. The patient continued immunosuppressive treatment and was alive at the end of the study.
Another patient underwent an emergency left lower lobectomy because of treatment-resistant hemoptysis. Microscopic histopathology of the removed lobe was previously reported.23 The patient later developed Budd-Chiari syndrome and died of hepatic failure (Patient 8, Table 5). He was also reported as a case resistant to anti-tumor necrosis factor.44
One patient who had enlarging pulmonary cavities underwent bullectomy and also repair of a bronchopleural fistula.
Decortication of the right lung was done in another patient who suffered from refractory pleural effusion caused by vena cava superior thrombosis. At the same time, a lung segment that seemed to be destructed was removed from the right lower lobe. Pathology revealed chronic nonspecific inflammation.
Figure 11 shows the Kaplan-Meier survival curve for mortality. A total of 12 (11 male, 1 female) (26%) patients died. Massive hemoptysis was the major cause of death (7/12, 58%) (see Table 5). No postmortem studies were available. The median follow-up was 8 months (IQR, 4–19 mo). The mortality rate was similar for those with PAA (26%; 9/34) and those with isolated PAT (23%; 3/13). The frequency of symptoms and clinical characteristics at first presentation did not differ between those who died and those who remained alive (data not shown), except for exercise-induced dyspnea, which tended to be more common among those who died (50% vs. 23%, p = 0.091). Among patients with PAA, patients with an aneurysm with the largest diameter ≥3 cm (8/15, 53%) were more likely to have a fatal outcome, compared to those with a diameter <3 cm (2/22, 9%), (p = 0.002). The mean sPAP was significantly higher (50 ± 13 mm Hg) in patients who died (n = 6) compared with patients who survived (n = 17) (35 ± 5 mm Hg) (p = 0.001). The frequency of intracardiac thrombi and pulmonary parenchymal changes did not differ between those patients who had died and those who remained alive (data not shown).
We surveyed 47 BD patients with pulmonary artery involvement from a single dedicated center for clinical, laboratory, and radiologic manifestations; treatment data; and survival. Only 4 patients (4/47) with extensive extrapulmonary large-vessel disease did not fulfill the ISG criteria for BD. There are reports of young men with similar disease manifestations. This has been named Hughes-Stovin syndrome, most likely an incomplete form of BD.7,15,17
The pulmonary symptoms we observed in the current survey were hemoptysis, dyspnea, fever, cough, and chest pain, in line with previous reports.2,18,22,24,50,51 Hemoptysis was massive in about 1/3 and was the cause of death in 7 (15%) patients. Fever is known to be rare in BD, and when present it is usually associated with vascular disease.6,25,27 Its high frequency (47%) in the current survey can be similarly explained. In the same way, a high acute-phase response is usually associated with vascular disease in BD.38 In this regard we emphasize that in 7 patients relapsing with parenchymal lesions and high acute-phase response, fever and cough of sudden onset, rather than the hemoptysis, were the dominant manifestations of the relapse.
It is to be noted that a considerable portion (91%) of patients had signs of active BD outside the lungs at the time of pulmonary artery involvement diagnosis. As pulmonary artery disease got better, most of the pulmonary as well as other BD symptoms and findings also showed a tendency to abate with time. Yet, despite the absence of any active pulmonary arterial disease, some patients still suffered from small bouts of hemoptysis, cough and dyspnea. The presence or absence of these symptoms was not associated with the presence of residual pulmonary artery occlusions in the CT scans obtained at the end of follow-up. It is to be noted that these symptoms were not very severe, as they did not significantly affect the patient’s daily routine.
Another piece of useful clinical information that emerged was that in about half (6/11 available) of the patients with isolated PAT, the initial chest radiograph was normal. This contrasted with what we observed in patients with PAA, where the initial chest radiographs were abnormal at initial presentation in 27 of 28 patients. These findings point out that whenever a patient with BD presents with respiratory symptoms, especially hemoptysis, a CT scan of the lungs is mandatory.
Certain well-recognized characteristics of pulmonary artery involvement in BD were apparent in the current survey. First, it occurs mostly in men.22,24,46,50,51 It appears a mean of 4 years after the disease onset,22,24,50 and is strongly associated with venous thrombosis elsewhere in the vascular tree22,24,46,50,51 but rarely with peripheral artery disease.35 PAA, and, as we now better realize, PAT, are mostly localized in the inferior lobes, tend to become frequently thrombosed, and can disappear completely with medical treatment.4,5,12,40,46,51
Although isolated PAT has been reported as part of pulmonary artery involvement in BD,51 the current survey is the first, to our knowledge, to compare in detail the clinical and radiographic features of either condition. Although both PAA and PAT are localized to the lower lobes, the number of vessels involved is significantly higher in PAA (see Table 3). It has been reported that patients with isolated PAT had less massive hemoptysis and mortality compared to patients with PAA.51 We also observed the same concerning massive hemoptysis, with the additional observation that lesser amounts of hemoptysis was also less frequent among patients with isolated PAT. However, differing from the results reported by Uzun et al,51 the mortality in the current study was similar in patients with PAA and patients with isolated PAT. It is also interesting to note that in 3 of 13 patients with isolated PAT, the lesions progressed to PAA, at the site of the PAT in all 3. Based on our observation that 1) both lesions present in a similar fashion, 2) they eventually carry the same prognosis, and 3) PAT lesions transform into PAA lesions in some patients, we speculate that PAT might be a forerunner of PAA.
Nodules and cavities have been reported before in association with pulmonary artery involvement in BD.2,13,18,20,45,46,50,51 However, we now learn that they could be as frequent as 85% at the time of presentation. The fact that only 17% of patients in whom PAA or PAT healed continued to have these nodules strongly suggests that they are associated with the active stages of pulmonary artery involvement.
One might speculate that nodules might signal the onset of pulmonary artery involvement, since in 15% they were present even before pulmonary artery disease was manifest. The majority of the nodules resolved quickly, whereas a few were slow changing and cavitating, suggesting 2 different pathologic processes. Quick-resolving nodular lesions may be similar to those seen in bronchiolitis obliterans organizing pneumonia (BOOP), as was demonstrated in 1 of our biopsies and in 2 case reports by other groups.21,39 Their presence during exacerbations, peripheral location, multiple and wandering pattern, association with pulmonary symptoms including fever, and rapid response to corticosteroids are features quite typical for BOOP.3,16,48 On the other hand, large, slow-changing pleura-based lesions and cavities themselves suggest necrosis or infarction, as shown in 3/5 of the biopsies in the current study. The fact that, in contrast to nodular lesions, most of the cavities developed during the follow-up, also supports this contention. It seems that histopathology of the nodular lesions shows great variety: from granuloma (as shown in the current study) to eosinophilic pneumonia or pulmonary hemorrhage (as demonstrated by Tunaci et al45). Similarly, as Uzun et al51 have suggested, some nodular lesions may well be related to their hypothesis of microscopic vascular disease.
Some patients also had ground-glass opacities, consolidations, pleural or pericardial lesions, and minor lymphadenopathy. Again, all these lesions were found in the active stages of the pulmonary artery involvement.
Intracardiac thrombosis is a relatively infrequent finding that has mainly been reported in association with PAA.12,26,30,36 In the current study, this complication was observed in about one-third of the patients either by echocardiography or by thoracic CT. In a cumulative review of 25 cases with intracardiac thrombosis, the intracardiac thrombi usually appeared soon after disease onset.36 They were mostly located in the right side of the heart, mainly in the ventricle, and responded well to immunosuppressives.26,30 This was also the case in the current study. Such thrombosis are tightly adherent to underlying endocardium or myocardium, and the histology shows an organizing thrombus containing inflammatory cell infiltrates, indicating that it develops secondary to an underlying endocarditis.36 The poor prognosis once ascribed36 to this lesion was most likely due to the coexistence of pulmonary artery involvement. In the current study, survival in patients with intracardiac thrombosis was not different from survival in patients without, and intracardiac lesions disappeared with effective treatment.
We observed sPAP elevation by echocardiography in 65% of the patients, similar to what has been described in a number of case reports.7,13,19,26,32,43 This was even more pronounced among those who had died, suggesting that pulmonary arterial hypertension could be a contributing factor to mortality. Since sPAP normalized in half of the patients during the follow-up, we think that sPAP usually increases during the acute phase of pulmonary artery disease.
An important and much debated issue in BD34 is whether thrombi in the lungs could be due to emboli, considering the high frequency of venous thrombosis in the lower extremities in these patients. However, this seems unlikely for a number of reasons: 1) Pulmonary embolism was not observed in any of the patients in the largest (to our knowledge) (n = 130) postmortem series in BD reported.33 2) While pulmonary arteritis is the primary underlying pathology in BD,13,23 arteritis is absent in garden variety pulmonary thromboembolism.42,53 3) Anticoagulation without immunosuppressive treatment is most probably ineffective in preventing relapses in BD, as shown in 2 retrospective studies and 1 review.1,29,50 And finally, 4) As we have observed in 6 patients in the current study, the mismatch areas in the V/Q scans persist during follow-up. This is different from what is seen in pulmonary thromboembolism, where resolution in time is often the rule.41,49 Similarly, Caglar et al8 observed persistence of the defects on the lung scintigraphy after 3 months in 5 BD patients with pulmonary involvement; they suggested that persistence of the defects on the V/Q scans would imply vasculitis. This last point has been also been brought up in a case report.14
It is to be noted that patients in the current study were intensively treated with immunosuppressives. Nevertheless, relapses were observed, even under CYP or AZA. Infliximab seems to be a good alternative in treating these patients.5 Due to reasons discussed above, we suggest that patients with isolated PAT would be better treated with immunosuppressive agents rather than anticoagulants.
Recently endovascular treatment has emerged as an important alternative for the treatment of PAA.9,32,37 Our embolization attempts were unsuccessful in 2 patients: As previously reported this procedure may have complications such as air cavity transformation of the aneurysmatic sac,10 and it may not be effective in large aneurysms. Nevertheless, it could be tried in treatment-resistant hemoptysis. We and others have previously reported that lobectomy had lethal results.11,43,47 However, in the current study lobectomy was done in 1 patient and segmentectomy in another, and both survived. We also observed that the patients with larger aneurysms had a guarded prognosis. Thus we think that aneurysms that do not regress under standard immunosuppressive treatment most probably have lost the vascular integrity, probably due to wall rupture and inflammatory invasion into adjacent anatomic structures. These aneurysms are more prone to rupture completely into a bronchus, leading to the immediate death of the patient. This suggests that in such well-selected patients, lobectomy might still be considered.
We now report that the mortality rate among our patients has not changed since 2004, at which time we had reported 23% mortality. This indicates that despite early recognition and treatment, about one-fourth of our BD patients with pulmonary artery involvement unfortunately still die from this complication.
There were some limitations to the current survey. As this was a retrospective study, radiologic imaging techniques were not standardized and uniform. Histopathologic data were sparse. No postmortem exams could be obtained.
In summary, the current survey highlights how pulmonary artery involvement in BD can present a complex clinical picture (Table 6). It not only involves the pulmonary arteries, but also affects the whole lung parenchyma and cardiac tissues. PAT is more important in this picture than we previously recognized. Many patients continue to have some respiratory symptoms even after the acute phase of pulmonary artery involvement has subsided and chest CT scans have normalized. Pulmonary arterial hypertension may be important to consider both in the acute stage and for the long term. Finally, and to our dismay, pulmonary artery involvement still carries a rather grave prognosis.
1. Ahn JK, Lee YS, Jeon CH, Koh EM, Cha HS. Treatment of venous thrombosis associated with Behcet’s disease: immunosuppressive therapy alone versus immunosuppressive therapy plus anticoagulation. Clin Rheumatol. 2008; 27: 201–205.
2. Ahn JM, Im JG, Ryoo JW, Kim SJ, Do YS, Choi YW, Oh YW, Yeon KM, Han MC. Thoracic manifestations of Behcet syndrome: radiographic and CT findings in nine patients. Radiology. 1995; 194: 199–203.
3. Akira M, Yamamoto S, Sakatani M. Bronchiolitis obliterans organizing pneumonia manifesting as multiple large nodules or masses. AJR Am J Roentgenol. 1998; 170: 291–295.
4. Aktogu S, Erer OF, Urpek G, Soy O, Tibet G. Multiple pulmonary arterial aneurysms in Behcet’s disease: clinical and radiologic remission after cyclophosphamide and corticosteroid therapy. Respiration. 2002; 69: 178–181.
5. Baki K, Villiger PM, Jenni D, Meyer T, Beer JH. Behcet’s disease with life-threatening haemoptoe and pulmonary aneurysms: complete remission after infliximab treatment. Ann Rheum Dis. 2006; 65: 1531–1532.
6. Bennesser-Alaoui H, Harmouche H, El Attar N, Tazi-Mezalek Z, Adnaoui M, Aouni M, Maaouni A. [Prolonged fever of unknown origin in Behcet’s disease. Pay attention to vessels!]. J Mal Vasc. 2008; 33: 239–241.
7. Bowman S, Honey M. Pulmonary arterial occlusions and aneurysms: a forme fruste of Behcet’s or Hughes-Stovin syndrome. Br Heart J. 1990; 63: 66–68.
8. Caglar M, Ergun E, Emri S. 99Tcm-MAA lung scintigraphy in patients with Behcet’s disease: its value and correlation with clinical course and other diagnostic modalities. Nucl Med Commun. 2000; 21: 171–179.
9. Cantasdemir M, Kantarci F, Mihmanli I, Akman C, Numan F, Islak C, Bozkurt AK. Emergency endovascular management of pulmonary artery aneurysms in Behcet’s disease: report of two cases and a review of the literature. Cardiovasc Intervent Radiol. 2002; 25: 533–537.
10. Cil BE, Turkbey B, Canyigit M, Kumbasar OO, Celik G, Demirkazik FB. Transformation of a ruptured giant pulmonary artery aneurysm into an air cavity after transcatheter embolization in a Behcet’s patient. Cardiovasc Intervent Radiol. 2006; 29: 151–154.
11. de Montpreville VT, Macchiarini P, Dartevelle PG, Dulmet EM. Large bilateral pulmonary artery aneurysms in Behcet’s disease: rupture of the contralateral lesion after aneurysmorrhaphy. Respiration. 1996; 63: 49–51.
12. Duzgun N, Anil C, Ozer F, Acican T. The disappearance of pulmonary artery aneurysms and intracardiac thrombus with immunosupressive treatment in a patient with Behcet’s disease. Clin Exp Rheumatol. 2002; 20 (4 Suppl 26): S56–S57.
13. Efthimiou J, Johnston C, Spiro SG, Turner-Warwick M. Pulmonary disease in Behcet’s syndrome. Q J Med. 1986; 58: 259–280.
14. el-Ramahi KM, Fawzy ME, Sieck JO, Vanhaleweyk G. Cardiac and pulmonary involvement in Behcet’s disease. Scand J Rheumatol. 1991; 20: 373–376.
15. Emad Y, Ragab Y, Shawki Ael-H, Gheita T, El-Marakbi A, Salama MH. Hughes-Stovin syndrome: is it incomplete Behcet’s? Report of two cases and review of the literature. Clin Rheumatol. 2007; 26: 1993–1996.
16. Epler GR, Colby TV, McLoud TC, Carrington CB, Gaensler EA. Bronchiolitis obliterans organizing pneumonia. N Engl J Med. 1985; 312: 152–158.
17. Erkan D, Yazici Y, Sanders A, Trost D, Yazici H. Is Hughes-Stovin syndrome Behcet’s disease? Clin Exp Rheumatol. 2004; 22 (4 Suppl 34): S64–S68.
18. Erkan F, Kiyan E, Tunaci A. Pulmonary complications of Behcet’s disease. Clin Chest Med. 2002; 23: 493–503.
18a. 18a. Geri G, Wechsler B, Huong DLT, Isnard R, Piette JC, Amoura Z, Resche-Rigon M, Cacoub P, Saadoun D. Spectrum of cardiac lesions in Behcet disease: a series of 52 patients and review of literature. Medicine (Baltimore). 2012; 91: 25–34.
19. Gibson RN, Morgan SH, Krausz T, Hughes GR. Pulmonary artery aneurysms in Behcet’s disease. Br J Radiol. 1985; 58: 79–82.
20. Grenier P, Bletry O, Cornud F, Godeau P, Nahum H. Pulmonary involvement in Behcet disease. AJR Am J Roentgenol. 1981; 137: 565–569.
21. Gul A, Yilmazbayhan D, Buyukbabani N, Lie JT, Tunaci M, Tunaci A, Inanc M, Ocal L, Aral O, Konice M. Organizing pneumonia associated with pulmonary artery aneurysms in Behcet’s disease. Rheumatology (Oxford). 1999; 38: 1285–1289.
22. Hamuryudan V, Er T, Seyahi E, Akman C, Tuzun H, Fresko I, Yurdakul S, Numan F, Yazici H. Pulmonary artery aneurysms in Behcet syndrome. Am J Med. 2004; 117: 867–870.
23. Hamuryudan V, Oz B, Tuzun H, Yazici H. The menacing pulmonary artery aneurysms of Behcet’s syndrome. Clin Exp Rheumatol. 2004; 22 (4 Suppl 34): S1–S3.
24. Hamuryudan V, Yurdakul S, Moral F, Numan F, Tuzun H, Tuzuner N, Mat C, Tuzun Y, Ozyazgan Y, Yazici H. Pulmonary arterial aneurysms in Behcet’s syndrome: a report of 24 cases. Br J Rheumatol. 1994; 33: 48–51.
25. Harmouche H, Maamar M, Sahnoune I, Tazi-Mezalek Z, Aouni M, Maaouni A. Fever revealing Behcet’s disease: two new cases. Eur J Intern Med. 2007; 18: 146–147.
26. Huong DL, Dolmazon C, De Zuttere D, Wechsler B, Godeau P, Piette JC. Complete recovery of right intraventricular thrombus and pulmonary arteritis in Behcet’s disease. Br J Rheumatol. 1997; 36: 130–132.
27. Inal S, Onec K, Uludag K, Derici UB, Sindel S. Prolonged fever: an atypical presentation of Behcet’s disease. Rheumatol Int. 2011; 31: 839–840. [Epub 2010 Apr 8].
28. International Study Group for Behcet’s Disease. Criteria for diagnosis of Behcet’s disease. Lancet. 1990; 335: 1078–1080.
29. Kahraman O, Celebi-Onder S, Kamali S, et al.. Long-term course of deep venous thrombosis in patients with Behcet’s disease. Arthritis Rheum. 2003; 48 (Suppl 9): S385.
30. Kaneko Y, Tanaka K, Yoshizawa A, Yasuoka H, Suwa A, Satoh T, Iwanaga S, Ogawa S, Ikeda Y, Hirakata M. Successful treatment of recurrent intracardiac thrombus in Behcet’s disease with immunosuppressive therapy. Clin Exp Rheumatol. 2005; 23: 885–887.
31. Kural-Seyahi E, Fresko I, Seyahi N, Ozyazgan Y, Mat C, Hamuryudan V, Yurdakul S, Yazici H. The long-term mortality and morbidity of Behcet syndrome: a 2-decade outcome survey of 387 patients followed at a dedicated center. Medicine (Baltimore). 2003; 82: 60–76.
32. Lacombe P, Qanadli SD, Jondeau G, Barre O, Mesurolle B, Mouas H, Igual J, Chagnon S, Dubourg O, Bourdarias JP. Treatment of hemoptysis in Behcet syndrome with pulmonary and bronchial embolization. J Vasc Interv Radiol. 1997; 8: 1043–1047.
33. Lakhanpal S, Tani K, Lie JT, Katoh K, Ishigatsubo Y, Ohokubo T. Pathologic features of Behcet’s syndrome: a review of Japanese autopsy registry data. Hum Pathol. 1985; 16: 790–795.
34. Mehta P, Laffan M, Haskard DO. Thrombosis and Behcet’s syndrome in non-endemic regions. Rheumatology (Oxford). 2010; 49: 2003–2004.
35. Melikoglu M, Ugurlu S, Tascilar K., Caglar E, Seyahi E, Hamuryudan V, Yurdakul V, Yazici H. Large vessel involvement in Behcet’s syndrome: a retrospective survey. Ann Rheum Dis. 2008; 67 (Suppl II): 67.
36. Mogulkoc N, Burgess MI, Bishop PW. Intracardiac thrombus in Behcet’s disease: a systematic review. Chest. 2000; 118: 479–487.
37. Mouas H, Lortholary O, Lacombe P, Cohen P, Bourezak SE, Deloche A, Azorin J, Guillevin L. Embolization of multiple pulmonary arterial aneurysms in Behcet’s disease. Scand J Rheumatol. 1996; 25: 58–60.
38. Muftuoglu AU, Yazici H, Yurdakul S, Tuzun Y, Pazarli H, Gungen G, Deniz S. Behcet’s disease. Relation of serum C-reactive protein and erythrocyte sedimentation rates to disease activity. Int J Dermatol. 1986; 25: 235–239.
39. Ning-Sheng L, Chun-Liang L, Ray-Sheng L. Bronchiolitis obliterans organizing pneumonia in a patient with Behcet’s disease. Scand J Rheumatol. 2004; 33: 437–440.
40. Numan F, Islak C, Berkmen T, Tuzun H, Cokyuksel O. Behcet disease: pulmonary arterial involvement in 15 cases. Radiology. 1994; 192: 465–468.
41. Palla A, Pazzagli M, Manganelli D, De Nitto P, Marini C, Rossi G, Mazzantini D, Giuntini C. Resolution of pulmonary embolism: effect of therapy and putative age of emboli. Respiration. 1997; 64: 50–53.
42. Piazza G, Goldhaber SZ. Chronic thromboembolic pulmonary hypertension. N Engl J Med. 2011; 364: 351–360.
43. Raz I, Okon E, Chajek-Shaul T. Pulmonary manifestations in Behcet’s syndrome. Chest. 1989; 95: 585–589.
43a. 43a. Saadoun D, Asli B, Wechsler B, Houman H, Geri G, Desseaux K, Piette JC, Huong DLT, Amoura Z, Ben Salem T, Cluzel P, Koskas F, Resche-Rigon M, Cacoub P. Long-term outcome of arterial lesions in Behcet disease: a series of 101 patients. Medicine (Baltimore). 2012; 91: 18–24.
44. Seyahi E, Hamuryudan V, Hatemi G, Melikoglu M, Celik S, Fresko I, Yurdakul S, Yazici H. Infliximab in the treatment of hepatic vein thrombosis (Budd-Chiari syndrome) in three patients with Behcet’s syndrome. Rheumatology (Oxford). 2007; 46: 1213–1214.
45. Tunaci A, Berkmen YM, Gokmen E. Thoracic involvement in Behcet’s disease: pathologic, clinical, and imaging features. AJR Am J Roentgenol. 1995; 164: 51–56.
46. Tunaci M, Ozkorkmaz B, Tunaci A, Gul A, Engin G, Acunas B. CT findings of pulmonary artery aneurysms during treatment for Behcet’s disease. AJR Am J Roentgenol. 1999; 172: 729–733.
47. Tuzun H, Hamuryudan V, Yildirim S, Besirli K, Yoruk Y, Yurdakul S, Yazici H. Surgical therapy of pulmonary arterial aneurysms in Behcet’s syndrome. Ann Thorac Surg. 1996; 61: 733–735.
48. Uner AH, Rozum-Slota B, Katzenstein AL. Bronchiolitis obliterans-organizing pneumonia (BOOP)-like variant of Wegener’s granulomatosis. A clinicopathologic study of 16 cases. Am J Surg Pathol. 1996; 20: 794–801.
49. The urokinase pulmonary embolism trial. A national cooperative study. Circulation. 1973; 47 (2 Suppl): II1–II108.
50. Uzun O, Akpolat T, Erkan L. Pulmonary vasculitis in Behcet disease: a cumulative analysis. Chest. 2005; 127: 2243–2253.
51. Uzun O, Erkan L, Akpolat I, Findik S, Atici AG, Akpolat T. Pulmonary involvement in Behcet’s disease. Respiration. 2008; 75: 310–321.
52. Value of the ventilation/perfusion scan in acute pulmonary embolism. Results of the prospective investigation of pulmonary embolism diagnosis (PIOPED). The PIOPED Investigators. JAMA. 1990; 263: 2753–2759.
53. Wagenvoort CA. Pathology of pulmonary thromboembolism. Chest. 1995; 107 (Suppl): 10S–17S.