The Shunt of It

Abstract

Pulmonary arteriovenous malformations (PAVMs) are rare and most often identified in patients with hereditary hemorrhagic telangiectasia (HHT). We describe a patient with severe hypoxemia and orthodeoxia with imaging findings consistent with PAVMs. Resected lung pathologic findings confirmed the presence of numerous microscopic vascular abnormalities within the right lower lobe that was consistent with diffuse pulmonary arteriovenous shunts. Family history was negative for HHT but was positive for pulmonary arterial hypertension (PAH) in two second-degree relatives. A vascular malformation gene panel was negative for genes that commonly are associated with HHT but identified a pathogenic variant in the gene encoding bone morphogenetic protein receptor-2 (BMPR2 p.Cys123∗). Pathogenic variants in BMPR2 are a well-known cause of hereditary PAH; there have been several reports to date of patients with PAVMs and PAH. However, this is the first patient to be reported with a pathogenic variant in BMPR2 to have PAVMs in isolation.

A 55-year-old woman with a medical history of treated OSA and a 20 pack-year smoking history presented for evaluation of progressive dyspnea, hypoxemia, and lightheadedness. Mild epistaxis occurred after initiation of supplemental oxygen. Family history, per patient report, was notable for pulmonary arterial hypertension (PAH) in her maternal aunt and grandmother (Fig 1)

Figure 1.

Family pedigree. Current patient listed as BMPR-2 variant with the arteriovenous malformations. No additional family members were available for evaluation or genetic testing. AVM = arteriovenous malformation; PAH = pulmonary arterial hypertension.

Physical examination revealed digital clubbing and oxygen saturation of 89% on 3 L/min nasal cannula, with an increase to 93% on 10 L/min nasal cannula. She had orthodeoxia desaturating to 85% on standing. Pulmonary function testing revealed normal spirometry and a reduced diffusion capacity of carbon monoxide of 15.01 mL/mm Hg/min (53% of predicted). Her right lower lobe pulmonary vessels were prominent on CT pulmonary angiogram (Fig 2). Transthoracic echocardiogram with agitated saline solution revealed intrapulmonary shunting with bubbles entering the left atrium after more than three cardiac cycles. Right heart catheterization was not consistent with a diagnosis of PAH (Table 1) or portal hypertension (hepatic venous pressure gradient of 2 mm Hg). Quantitative lung perfusion scintigraphy revealed a 18.9% right-to-left shunt concerning for pulmonary arteriovenous malformations (PAVMs). Conventional pulmonary angiography revealed early draining of the right lower lobe pulmonary veins that is consistent with microvascular PAVMs (Fig 3, Video 1).

Figure 2.

A and B, Prominence of right lower lobe pulmonary vessels on A, axial and B, coronal slices on CT pulmonary angiogram.

Table 1.

Right Heart Catheterization Values

Systolic Pulmonary Artery Pressure, mm Hg	Diastolic Pulmonary Artery Pressure, mm Hg	Mean Pulmonary Artery Pressure, mm Hg	Pulmonary Capillary Wedge Pressure, mm Hg	Cardiac Output, L/min	Cardiac Index, L/min/m2	Pulmonary Vascular Resistance, Wood Units
26	9	18	12	5.92	2.99	1.01

Figure 3.

Conventional pulmonary angiogram shows increased peripheral vasculature in the right lower lobe and early draining of enlarged lower lobe pulmonary veins. Findings were concerning for arteriovenous malformations and microvascular shunting.

Because innumerable PAVMs were present throughout the entire right lower lobe, they were not amenable to embolization. Therefore, despite the increased potential of the development of pulmonary hypertension by reducing the pulmonary vascular bed,¹ a right lower lobectomy was performed to treat severe arterial hypoxemia. Pathologic evidence revealed abnormally dilated, variably sized, and somewhat disorganized vessels that were seen throughout the lobules (Fig 4A), which is a histologic pattern previously described in arteriovenous shunting.² Elastin stains identified these vessels as having delicate elastic walls consistent with veins (Fig 4B). Occasional arteries were noted near the disorganized dilated vessels within the lobules (Fig 5), and hypertrophic arteriopathy of medium size arteries was also noted (Fig 6), compatible with localized increased pressures of nearby aneurysmal veins. No evidence of plexogenic arteriopathy was identified. After surgery, the patient’s oxygen requirements decreased, although she still required 1 L/min supplementation overnight.

Figure 4.

A and B, Histologic findings of pulmonary vascular abnormalities. A, Microscopic section of the right lower lobe shows multiple engorged, variably dilated thin vessels throughout the lobules (hematoxylin and eosin stain; original magnification, ×40). B, Elastin stain highlights a delicate elastic layer in these vessels, compatible with veins (original magnification, ×200).

Figure 5.

Histologic finding of pulmonary vascular abnormalities. Elastin stain shows larger arteries near disorganized smaller veins in periphery of lobules (original magnification, ×100).

Figure 6.

Microscopic section of right lower lobe shows focal hypertrophic arteriopathy with markedly thickened pulmonary arteries (hematoxylin and eosin stain; original magnification, ×100).

A vascular malformation gene panel (genes tested in ARUP Laboratories vascular malformation panel: ACVRL1; AKT1; BMPR2; CCBE1; CCM2; EIF2AK4; ELMO2; ENG; EPHB4; FAT4; FLT4; FOXC2; GATA2; GDF2; GJC2; GLMN; KCNK3; KRIT1; PDCD10; PIEZO1; PTEN; RASA1; SMAD4; SMAD9; SOX18; STAMBP, TEK; VEGFC) revealed a previously unreported pathogenic variant in bone morphogenetic protein receptor-2 (BMPR2 c.369delT;p.Cys123∗) and was otherwise negative, including for hereditary hemorrhagic telangiectasias (HHT) genes. This variant induces an early termination codon and is predicted to result in either a truncated protein or no protein product.

Discussion

PAVMs are uncommon,^{3, 4, 5} with 70% to 95% of PAVMs occurring in patients with HHT.^{4,6, 7, 8, 9} Although non-HHT-related PAVMs can occur in congenital heart disease, hepatopulmonary syndrome, and after thoracic surgery, trauma, and pulmonary infections, most are idiopathic and isolated.¹⁰ The majority (75% to 90%) of PAVMs are simple with a single feeding artery,^5,8,9,11 frequently identified in the lower lobes and unilateral in nature.^6,12 The remaining 10% to 25% are either complex PAVMs with at least two feeding arteries or diffuse PAVMs with many feeding arteries.^5,8,10,11 Whereas patients with HHT commonly have multiple simple PAVMs, patients with no HHT more frequently have a single PAVM.^9,10

HHT is an autosomal-dominant disorder characterized by recurrent epistaxis, telangiectasia, and arteriovenous malformations in characteristic locations (lips, fingers, oral cavity, lungs, GI tract, and CNS).¹³ However, because of the rarity of PAVMs, their presence, even in isolation, can suggest a diagnosis of HHT.¹⁴

BMPR2 and the HHT genes (ACVRL1, ENG, and SMAD4) are all part of the transforming growth factor-beta signaling pathway.¹⁵ Pathogenic variants in BMPR2 are associated overwhelmingly with heritable PAH,¹⁶ though none have been reported to date in a patient who has been diagnosed with HHT. There are, however, several descriptions of patients with both PAVMs and PAH in the setting of pathogenic BMPR2 variants.^{17, 18, 19} Conversely, the development of PAH in patients with HHT is an infrequent, yet recognized, entity that occurs more often in patients with mutations in ENG or ACVRL1.²⁰

This case, in which PAVMs without PAH is reported in a patient for the first time, suggests that assessment for pathogenic variants in BMPR2 may be helpful to make a molecular diagnosis in patients with PAVM. Ordering an expanded vascular malformation panel should be considered, particularly when evidence of HHT from a patient’s clinical evaluation and family history is absent or minimal. This approach optimizes the chance for diagnosis in the patient and for family members to receive genetic counseling and testing to facilitate screening for PAH if a BMPR2 variant is identified.

Funding/Support

This study was funded by the National Institutes of Health under Ruth L. Kirschstein National Research Service Award 5T32HL105321 from the NIH.

Financial/Nonfinancial Disclosures

None declared.