Lymphangioleiomyomatosis manifesting as refractory chylothorax and chyloperitoneum

Abstract

This is a rare case of sporadic lymphangioleiomyomatosis (S-LAM) manifesting as refractory chylothorax and chyloperitoneum. A middle-aged woman with unremarkable medical history presented with respiratory failure, abdominal distension and anasarca. She was found to have high-output chylous effusion that required chest tube drainage, as well as chylous ascites. Notably initial chest and abdominal CT did not reveal characteristic pulmonary cysts or the presence of angiomyolipomas suggestive of LAM. An extensive oncologic and infectious work-up was undertaken with negative findings. The chylous effusion was persistent and refractory to thoracic duct embolization, total parenteral nutrition with octreotide, and talc pleurodesis. Diagnosis of S-LAM was confirmed after repeat chest CT showed subtle pulmonary cystic changes, and serum vascular endothelial growth factor-D level was found to be elevated at 834 pg/mL. Patient was started on sirolimus therapy, but lost to follow-up after hospital discharge. Patient died approximately 1 year later.

Background

Lymphangioleiomyomatosis (LAM) is a neoplastic disorder of smooth muscle-like cells proliferating and infiltrating into multiple organs, and can result in a range of clinical manifestations including diffuse pulmonary cysts, spontaneous pneumothorax, lymphatic abnormalities, and benign tumours such as angiomyolipomas and lymphangioleiomyomas.¹ It can occur in association with tuberous sclerosis complex (TSC-LAM), or in a sporadic form (S-LAM) that most commonly afflicts young premenopausal women. Given its rarity and non-specific presentation, the diagnosis of LAM can be challenging, and current guidelines rely primarily on the presence of pulmonary cysts on CT imaging to arouse clinical suspicion.^{2 3} Here we report a case of atypical S-LAM manifesting as refractory chylothorax and chyloperitoneum, with minimal pulmonary cystic changes. Our case illustrates the hazard of relying on imaging in the initial evaluation for suspected LAM, and suggests further refinements to the diagnostic algorithm are needed in order to fully encapsulate the natural history of this disease.

Case presentation

A Caucasian woman in her mid-40s initially presented to an outside community hospital with progressive dyspnea on exertion developing over several days, now worsening to dyspnea at rest; she also noted new onset abdominal distension and lower extremity oedema. Medical history includes chart history of chronic obstructive pulmonary disease (no previous pulmonary function test available), hypertension and a remote history of deep vein thrombosis that was previously treated with warfarin therapy. She was a former smoker, half-pack per day for twenty years, quit approximately 1 year prior to presentation. She was nulliparous; social and family history were otherwise unremarkable. Work-up at the outside hospital included chest X-ray which showed bilateral pleural effusions, with thoracentesis yielding milky fluid with negative cultures and cytology. Chest and abdominal CT reportedly demonstrated bilateral pleural effusions and ascites; no pulmonary embolism was noted. Axillary and mediastinal lymph node biopsies were negative for malignancy. She was placed on low-fat diet and aggressively diuresed with parenteral diuretics without improvement, and subsequently developed oliguric acute kidney injury. Due to rapid re-accumulation of pleural effusion, a left-side chest tube was placed for drainage. She was then transferred to our medical centre for further evaluation and management of her chylothorax.

On initial examination, she had an obese body habitus and was ill-appearing, requiring supplemental oxygen at 3 L/min to maintain oxygen saturation. Auscultation demonstrated poor air movements bilaterally; a left-side chest-tube drained milky fluid, with approximately 2.1 L of output in the past 24 hours. Abdominal examination showed mild distension and diffuse anasarca. She also had bilateral symmetric 3+pitting oedema in her lower extremities. Her remaining exams were otherwise normal.

Investigations

Initial chemistry was significant for elevated creatinine of 5.14 mg/dL (baseline 1.0 mg/dL) and albumin 2.0 g/dL, but otherwise within normal limits. Complete blood count showed normocytic anaemia with haemoglobin 83 g/L, mild thrombocytosis of 471 000/mm³; no leucocytosis. Liver function and lipid panel were unremarkable. Urine chorionic gonadotropin was negative. HIV, hepatitis panel, urine histoplasma antigen, urine culture and blood cultures were negative. Chest X-ray showed small left pleural effusion with left thoracostomy tube in place, and moderate right pleural effusion that appeared to be free-flowing. Transthoracic echocardiogram showed hyperdynamic ejection fraction 71%, mild concentric left ventricular hypertrophy and small-to-moderate pericardial effusion without haemodynamic effect. Given patient’s renal dysfunction on presentation, chest CT without contrast was initially performed and showed bilateral pleural effusions and nodular ground glass opacities in the right lower lobe, without obvious injury or obstruction of the thoracic duct; no suspicious cystic changes were noted. A repeat chest CT with contrast was later performed, again demonstrated bilateral pleural effusions, without any pulmonary emboli. Abdominal CT showed small volume ascites, and enlarged lobular uterus with internal calcifications thought likely fibroids. Pelvic ultrasound was performed and confirmed multiple fibroids with scattered calcifications, without other gynecologic abnormalities. Mammogram and whole-body fluorodeoxyglucose-positron emission tomography (FDG-PET) were also performed and were negative for malignancy. Serum and urine protein electrophoresis and immunofixation were negative for monoclonal proteins.

Analysis of pleural fluid collected from the left-side chest tube was significant for triglyceride of 615 mg/dL. Right-side therapeutic thoracentesis was performed and yielded 1.7 L of serous fluid, glucose 100 mg/dL, protein <2.0 g/dL, triglyceride 32 mg/dL, lactate dehydrogenase 58 mg/dL (pleural-to-serum ratio 0.3), cell count with 227 nucleated cells (40% neutrophils, 46% monocytes, 7% lymphocytes, 3% eosinophils and 4% mesothelial cells). All pleural fluid cultures and stains for aerobic, anaerobic, fungal and acid-fast organisms were negative. Pleural fluid cytology and flow cytometry were negative for malignancy and B-cell clonality. Paracentesis yielded 5.0 L of milky fluid, with albumin <1.0 mg/dL, triglyceride 782 mg/dL. All peritoneal fluid cultures including aerobic, anaerobic and fungal organisms were negative. Peritoneal fluid cytology was negative for malignancy. Bronchoalveolar lavage showed 170 nucleated cells (82% monocytes, 15% neutrophils, 3% lymphocytes); microbiologic studies including aerobic culture, acid-fast bacilli culture, respiratory virus panel, herpes simplex virus, cytomegalovirus, legionella and pneumocystis were negative. Serum level of vascular endothelial growth factor-D (VEGF-D) was elevated at 834 pg/mL. Brain MRI without contrast showed no evidence of tuberous sclerosis.

Differential diagnosis

The aetiology of chylothorax can be broadly divided into traumatic and non-traumatic. Among traumatic causes, prior surgical intervention is the predominant mode of injury, accounting for 25%–48% of all cases. Among non-traumatic causes, malignancy is the leading culprit, accounting for 17%–46% of all cases, with lymphomas being the most frequently implicated.^{4 5} Given our patient did not have any recent history of surgery or traumatic injury, we initially considered malignancy as the most likely aetiology for her chylothorax and chyloperitoneum. Notably, given the lack of characteristic diffuse cystic changes in the lungs, as well as the absence of angiomyolipomas or history of pneumothorax, LAM was considered to be less likely. Although uterine masses were identified on initial imaging, they were thought to be more likely fibroids rather than uterine LAM lesions; uterine biopsy was discussed but ultimately it was thought that it would not be of substantial aid in the diagnosis.

We undertook an aggressive oncologic work-up including repeat chest and abdominal CT, pelvic ultrasound, mammography, pleural fluid cytology and flow cytometry, peritoneal fluid cytology, FDG-PET and protein electrophoresis and immunofixation, all of which were negative. Although infection is not commonly associated with the development of chylothorax, we also undertook extensive infectious work-up and ruled-out any superimposed process that may contribute to the patient’s clinical presentation.

While these studies were ongoing, the patient continued to have high-output chyle leak from the left-side chest-tube approximately 1.0–1.8 L/day. Multiple therapeutic thoracenteses were also performed for the right-side effusion with rapid re-accumulation, eventually necessitating placement of a right-side chest tube as well. Patient continued to exhibit diffuse anasarca, thought likely due to continuous protein loss in the setting of chyle leak, resulting in low plasma oncotic pressure and fluid extravasation. Due to concern for worsening respiratory failure, our immediate priority was to limit the chylous effusion. Patient was continued on low-fat diet to reduce chyle production.⁶ Based on previous case reports of idiopathic chylothorax associated with simvastatin use,^{7 8} her home medication of lovastatin was discontinued, although this did not impact her clinical course. Total parenteral nutrition (TPN) has been shown to reduce chyle flow and production, as well as promote spontaneous closure of thoracic duct injury, and is usually the therapy of choice in patients not requiring immediate invasive intervention.^{6 9} However given the very high-output chylous effusion and tenuous respiratory status, as well as patient’s resistance to extended TPN trial, we elected for immediate thoracic duct embolisation (TDE) as the initial approach, which has a reported success rate of up to 90%.¹⁰ Lymphangiography was performed and showed high-flow through an enlarged cisterna chyli (>1 cm); the thoracic duct was then embolised with Nester coils and N-butyl cyanoacrylate. However, patient continued to have high-output chylothorax when challenged with high-fat diet, and a second TDE was performed but still without improvement. Patient then agreed to be made nil per os (NPO) and was started on TPN. Octreotide has been shown previously to be an effective adjunctive therapy in the management of low-output chylothorax,¹¹ and this was started in conjunction with TPN. Unfortunately this approach also proved to be unsuccessful, and as a last resort, left-side video-assisted thoracoscopy with talc pleurodesis followed by repeat bedside talc pleurodesis was performed to obliterate the pleural space. Chest tube output decreased significantly afterwards to 100–200 mL/day, and bilateral chest tubes were removed. Shortly afterwards however, patient developed respiratory failure and required intubation and ventilatory support in the intensive care unit. Repeat non-contrast chest and abdominal CT showed recurrence of bilateral pleural effusions and ascites, as well as increasing pericardial effusion. Although not noted in the final radiology report, independent review by the pulmonology consulting service was concerning for subtle cystic parenchymal changes (figure 1). Patient serum sample was sent to Cincinnati Children’s Hospital Medical Centre for VEGF-D testing, and was highly-elevated at 834 pg/mL, consistent with the diagnosis of LAM. Given her compatible clinical presentation and imaging finding, tissue biopsy was deemed unnecessary. Brain MRI was performed and did not show any lesions suspicious for TSC.

Figure 1.

Non-contrast chest CT demonstrating sporadic pulmonary cysts.

Treatment

Patient was stabilised in the intensive care unit, and required replacement of left-side chest tube. Paracentesis was also performed for large volume ascites in order to improve diaphragmatic function. Respiratory status gradually improved, and patient was extubated and chest tube was removed without complications. Patient was started on sirolimus therapy, Bactrim prophylaxis, and counselled on low-fat diet. Patient was discharged to a nursing facility due to deconditioning as a result of prolonged hospitalisation.

Outcome and follow-up

Patient subsequently lost to follow-up after hospital discharge, and had no further records in our system. According to family, patient subsequently had two additional hospitalizations at outside facilities, and died at home on hospice care approximately 1 year later.

Discussion

LAM is a rare neoplastic disease characterised by smooth muscle-like cells proliferating and infiltrating into the lung parenchyma, airways, lymphatics, and blood vessels, resulting in multisystem manifestations. The defining characteristic of LAM disease is the presence of bilateral diffuse pulmonary cysts on CT imaging; extrapulmonary features include angiomyolipomas most commonly arising from the kidneys, lymphangioleiomyomas that can present as retroperitoneal mass, aberrations of lymphatic flow resulting in chyle leaks into the thorax and abdomen, and uterine LAM lesions. The clinical presentation of LAM is usually non-specific, with the most frequent chief complaint being progressive dyspnea.¹² It can be frequently misdiagnosed as asthma or chronic obstructive pulmonary disease (COPD), as the clinical manifestations overlap, and pulmonary function test frequently shows obstructive defect¹³; indeed, we suspect the previous diagnosis of COPD in our patient may have reflected early manifestations of LAM. However, a history of spontaneous pneumothorax can be an important diagnostic clue, as it precedes the diagnosis of LAM in 82% of cases,¹⁴ and in one cross-sectional study LAM patients with a history of pneumothorax had on average four prior episodes.¹²

The current gold-standard of diagnosis remains lung biopsy followed by staining with the monoclonal antibody HMB-45 that is highly specific for LAM cells¹⁵; however this is increasingly supplanted by VEGF-D testing, with a threshold value of >800 pg/mL achieving 71% sensitivity and 100% specificity.^16–18 The latest American Thoracic Society/Japanese Respiratory Society Clinical Practice Guideline recommended serum VEGF-D testing as the initial laboratory test after chest CT is obtained.³ It should be noted that this test is not yet widely accessible; in the United States it is currently only available at the Translational Trials Development and Support Lab at Cincinnati Children’s Hospital Medical Center.

The atypical presentation of our patient provided unique challenges in the diagnosis of LAM. In particular the lack of significant pulmonary cystic changes on imaging lowered our initial clinical suspicion. Indeed, existing guidelines largely focus on the use of chest CT to detect the presence of diffuse pulmonary cysts, either as the main diagnostic criterium,² or as the initial diagnostic test of choice.³ However it is increasingly recognised that LAM is a neoplastic disease, and can have widely variable clinical presentation based on the organ system affected. Our case illustrates the pitfall of over-reliance on lung imaging for the initial diagnostic consideration of LAM. As the cellular and molecular mechanisms in the pathogenesis of LAM are further elucidated, it would be beneficial to refine the definition and diagnostic algorithm of LAM in the proper context of a neoplastic multisystem disorder, rather than a cystic lung disease with associated extrapulmonary features.

Learning points.

• Lymphangioleiomyomatosis (LAM) is a neoplastic multisystem disorder and clinical features other than cystic lung disease on imaging may be more prominent on presentation.

• After chest CT, serum vascular endothelial growth factor-D level is the initial laboratory test of choice in cases of suspected LAM.

• In the absence of malignancy, the diagnosis of LAM should be considered in cases of non-traumatic chylothorax and/or chyloperitoneum.