SciELO - Scientific Electronic Library Online

vol.54 issue4P ET CO2 and SpO2 allow adequate ventilatory adjustment in morbidly obese patientsAnterior "3-in-1" blockade: partial, total or overdimensioned block? Correlation between anatomy, clinic and radio images author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links


Revista Brasileira de Anestesiologia

Print version ISSN 0034-7094On-line version ISSN 1806-907X

Rev. Bras. Anestesiol. vol.54 no.4 Campinas July/Aug. 2004 



Hypoxemia by pulmonary arteriovenous fistulae in childhood. Case report*


Hipoxemia por fístulas artério-venosas pulmonares en niño. Relato de caso



Aleksandra Paula Lima, M.D.I; José Lopes de Caíres, M.D.II; Dijair Gomes de Pontes, M.D.III; Fábio Biscegli Jatene, M.D.IV; José Otávio Costa Auler Júnior, TSA, M.D.V; Maria José Carvalho Carmona, TSA, M.D.VI

IME em Anestesiologia do HCFMUSP
IIAssistente do Serviço de Anestesia do Instituto do Coração do HCFMUSP
IIIEstagiário de Especialização em Anestesiologia e Pós-Operatório de Cirurgia Cardiovascular e Torácica do Instituto do Coração do HCFMUSP
IVProfessor Associado do Departamento de Cardiopneumologia da Faculdade de Medicina da Universidade de São Paulo. Diretor do Serviço de Cirurgia do Tórax do ICHC do Instituto do Coração do HCFMUSP
VProfessor Titular da Disciplina de Anestesiologia da FMUSP. Diretor do Serviço de Anestesiologia e Terapia Intensiva do Instituto do Coração do HC-FMUSP
VIProfessora Doutora da FMUSP. Médica Supervisora do Serviço de Anestesiologia do Instituto do Coração do HCFMUSP





BACKGROUND AND OBJECTIVES: Pulmonary arteriovenous fistulae (PAVF) should be investigated in patients with cyanosis of unknown cause. This is a case of cyanosis in a child submitted to pulmonary lobectomy with PAVF.
CASE REPORT: Male patient, 3 years old, with history of cyanosis without dyspnea since 8 months of age. He presented significant cyanosis, finger clubbing and normal heart auscultation. Chest X-ray showed condensation in the upper half of the left lung. Normal echocardiography and ECG. PaO2 = 28 mmHg in room air and PaO2 = 31.5 mmHg with nasal O2 catheter. Pulmonary arteriovenous fistula was diagnosed through magnetic resonance imaging, with no possibility of embolization. Patient was submitted to upper left lobectomy under general anesthesia associated to spinal anesthesia with morphine and bupivacaine. Selective tracheal tube for the right lung was inserted for monopulmonary ventilation. SaO2 was 59% in room air; at operating room admission 69% (FiO2 = 1.0); after general anesthesia induction 65% (FiO2 = 1.0); during monopulmonary ventilation 58% (FiO2 = 1.0), after lobectomy 98% (FiO2 = 1.0) and 98% at the end of the surgery (FiO2 = 0.6). Extubation was performed one hour after surgery completion. As from the fifth postoperative day patient started presenting progressive SpO2 decrease down to 83% due to increase of another PAVF, which was embolized under general anesthesia. Patient was discharged on the 15th PO day.
CONCLUSIONS: In this case, there was poor preoperative response to oxygen therapy, and patient has immediately improved after surgical treatment. However, after surgical resection of the largest PAVF, another fistula increased, thus leading to decrease in postoperative SpO2, reverted by embolization. As the child grows, other pulmonary fistula or fistulae in other organs may be diagnosed, indicating Rendu-Osler-Weber Syndrome.

Key Words: ANESTHESIA, Pediatric, DISEASES: Pulmonary: arteriovenous fistula


JUSTIFICATIVA Y OBJETIVOS: La presencia de fístulas artério-venosas pulmonares (FAVP) debe ser investigada en pacientes con cianosis sin causa esclarecida. El objetivo es relatar un caso de lobectomía pulmonar en niño portador de FAVP.
RELATO DO CASO: Paciente del sexo masculino, 3 años, con historia de cianosis sin disnea desde los 8 meses, con baqueteamento digital y ausculta cardíaca normal. ECG y ecocardiografía eran normales y la radiografía de tórax mostró condensación en la mitad superior del pulmón izquierdo. La gasometría arterial mostró PaO2 de 28 mmHg, en aire ambiente, y PaO2 de 31,5 mmHg con catéter nasal de O2. Fue hecho diagnóstico de FAVP a través de resonancia nuclear magnética, sin posibilidad de embolización. El paciente fue sometido a lobectomía superior izquierda sobre anestesia general asociada a anestesia subaracnóidea con morfina y bupivacaína. Fue realizada ventilación monopulmonar con introducción de tubo traqueal selectivo para el pulmón derecho. La SaO2 era: en aire ambiente de 59%; la admisión a la sala de operaciones, 69% (FiO2 = 1,0); después de inducción de la anestesia general, 65% (FiO2 = 1,0); durante la ventilación monopulmonar, 58% (FiO2 = 1,0); después de lobectomía y 98% (FiO2 = 0,6) al final da cirugía, con extubación traqueal una hora después de admisión en la UTI. A partir del 5º día de pós-operatorio, comenzó a presentar progresiva disminución de la SpO2 (hasta 83%) debido a aumento de otra FAVP, tratada con embolización bajo anestesia general. La alta hospitalar ocurrió en el 15º de pós-operatorio.
CONCLUSIONES: En este caso, había refractariedad a la suplementación con O2 y hubo mejoría inmediata con el tratamiento quirúrgico. Entretanto, después de resección de la FAVP de mayor tamaño ocurrió aumento de otra fístula, que contribuyó para la disminución de la SpO2 en el pós-operatorio, revertida con la embolización. Con el crecimiento del niño, otras fístulas pulmonares o en otros órganos podrán surgir, caracterizando la síndrome de Rendu-Osler-Weber.




Pulmonary arteriovenous fistula is an uncommon pulmonary vascular malformation which should be investigated in patients with cyanosis of unknown cause, and may be associated to Rendu-Osler-Weber syndrome. Although most patients are asymptomatic, PAVF may promote dyspnea by right-left shunt1. In general it is presented with cyanosis and dyspnea in infants or, later on, with cyanosis and/or abnormal chest X-rays2. This report aimed at describing a patient referred for persistent cyanosis investigation. PAVF was diagnosed and pulmonary lobectomy was performed.



Male patient, 3 years old, 13.5 kg born from vaginal delivery (BW = 3.550 g) without intercurrences. Patient's history revealed cyanosis since 8 months of age, without dyspnea and two admissions with diagnostic hypothesis of pneumonia. Radiological imaging showing upper left lobe condensation, has not improved after therapy with antibiotics. Patient persisted with cyanosis and congenital heart disease has been ruled out. At physical evaluation, patient presented with cyanosis (3+/3+), normal heart auscultation, clock glass nails, respiratory auscultation with abolishment of normal sounds on left pulmonary apex. Chest X-rays has shown increased right atrium and superior vena cava, aortic arch to the left, normal pulmonary vascular image and condensation in the upper half of left lung. ECG was normal. Echocardiography has shown large amount of left atrium bubbles after intravenous contrast injection in right arm. Arterial blood gas analysis in room air revealed pH 7.367, PaCO2 34.4 mmHg, PaO2 28 mmHg, HCO3 19.5 mEq/L, BE -4.7 mEq/L, SaO2 50.6%. With O2 catheter, values were pH 7.361, PaCO2 37.9 mmHg, PaO2 31.5 mmHg, HCO3 21 mEq/L, BE -3.5 mEq/L, SaO2 57.9%. Patient was referred to INCOR - HCFMUSP where MRI and pulmonary arteriography were performed and revealed pulmonary arteriovenous fistulae (PAVF) with no possibility of embolization. We decided for upper left lobectomy. Monitoring consisted of cardioscope, pulse oximetry, capnographer, CVP, diuresis, invasive blood pressure and serial blood gas analysis. Peripheral venoclysis with double-lumen catheter was performed in left subclavian vein. General anesthesia was induced with midazolam (4 mg), fentanyl (40 µg) and atracurium (10 mg), and was maintained with propofol (600 mg), atracurium (55 mg) and fentanyl (100 µg). Tracheal intubation was achieved with a 5 mm cuffed tube. Spinal anesthesia was associated with 75 µg morphine and 7.5 mg hyperbaric bupivacaine. Selective tracheal tube for the right lung was inserted for monopulmonary ventilation. SaO2 was 59% at operating room admission, 69% after general anesthetic induction, 65% during monopulmonary ventilation, 58% after lobectomy and 98% at surgery completion (Table I). Surgery went on without intercurrences and patient was extubated one hour after surgery completion. After PACU discharge and referral to the ward, patient started to present progressive SpO2 decrease (up to 83% in 11th PO), being diagnosed another pulmonary arteriovenous fistula in the right upper lobe (Figure 1), submitted to embolization under general anesthesia. Pathological exam of the specimen (Figure 2) has revealed pulmonary parenchyma partially replaced by an entangling of anastomosing vascular canals of different sizes, compatible with PAVF and suspicion of Rendu-Osler-Weber syndrome. Patient was discharged in the 15th postoperative day.



Afferent pulmonary circulation supply normally comes from the pulmonary artery, but may also be derived from systemic circulation by bronchial or intercostal arteries or by a aortic branch. Venous drainage is often done to pulmonary veins but, very seldom, there may be direct communication with left atrium. Most pulmonary arteriovenous fistulae are sub-pleural or located on the external third of the pulmonary parenchyma. Some times, these are large cavernous fistulae tending to infiltrate bronchial wall, which may result in rupture with massive hemorrhage.

PAVF incidence is 2 to 3 per 100 thousand people. Other names for this disease include benign cavernous hemangioma, pulmonary arteriovenous angiomatosis, lung hamartomatous angioma, arteriovenous aneurysm and pulmonary arteriovenous malformation. Female gender is more often affected. Multiple arteriovenous fistulae are found in 33% to 50% of patients, being bilateral in 8% to 20% of cases. Pick et al. have reported more than 80% of cases as congenital, and from these, almost half were associated to Rendu-Osler-Weber syndrome1. Rendu-Osler-Weber is a dominant autosomal syndrome related to chromosome 9q33-34 and chromosome 12q. The incidence of PAVF is particularly high in patients with genetic binding to chromosome 9q343 (29.2% versus 2.9% in patients without this binding)4. The disease is characterized by arteriovenous malformations and skin and mucosal telangiectasis, especially in upper respiratory and GI tract3. PAVF is seen in 27% of patients with Rendu-Osler-Weber syndrome and is associated to paroxysmal brain embolism rate in up to 36%. All Rendu-Osler-Weber syndrome patients are at some risk for PAVF, which makes important the evaluation of all of them, at least by the Hughes method (pulse oximetry in HDD and supine position), in addition to chest X-rays4.

PAVF in Rendu-Osler-Weber syndrome patients is in general diagnosed in the third or fourth decade of life because patients refer dyspnea, hemoptysis and, some times, epistaxis and hematemesis in the presence of upper respiratory and GI tract arteriovenous fistulae3. In children, most cases are diagnosed in infants or pre-school age and fistulae are located in lower lobes. Cyanosis and dyspnea are frequently found with systolic or continuous murmur2, most important during inspiration3, in the corresponding chest region in 60% of cases. ECG is frequently described as normal but there may be left chambers overload2. Arterial blood gas analysis in room air after pure oxygen inhalation during 20 minutes allow for separating right-left cardiac shunt from other ventilation-perfusion abnormalities, such as PAVF. In the first case, breathing pure oxygen may lead to high PaO2, differently from PAVF in which there will be just a slight improvement3.

PAVF may be secondary to other diseases, such as trauma, infection (actinomycosis, schistosomosis), chronic liver cirrhosis, mitral stenosis, metastatic carcinoma and systemic amyloidosis. Pregnancy may worsen PAVF patients status1. Differential diagnosis should be obtained by abnormal connection between left pulmonary artery and left atrium, which present similar pathophysiology2. Risk factors for the development of symptoms include young age, fistula above 2 cm diameter and Rendu-Osler-Weber syndrome. Symptoms include fatigue, dyspnea at exercise and palpitations. Classic triad is dyspnea at exercise, cyanosis and clubbed fingers in 30% of adults. Depending on right-left shunt degree, hypoxemia may be refractory to O2 therapy in 80% of cases. Moreover, supine position may be associated to additional arterial oxygen content decrease (Hughes method). Central neurological symptoms include headache, vertigo, paresis, syncope and dysphasia in up to one third of patients. Major central nervous system complications include seizures, abscesses, transient brain ischemia and stroke.

For PAVF diagnosis, chest X-rays will show circumscribed lesion with lobed density, mostly located in lower lobes1. Pulmonary opacification at plain chest X-rays and left chambers overload at ECG and echocardiography are data suggesting lesion2. Helix CT-scan may elucidate diagnosis in 95% of cases and may also be used as follow-up for lesion growth and check embolization results. Intravenous echocardiography contrasted with saline may be used both for diagnosis and PAVF screening in families with Rendu-Osler-Weber syndrome1. Sometimes, when PAVF is microscopic and diffuse, only perfusion pulmonary scintigraphy and blood gas analysis may suggest the diagnosis. PAVF should be considered in patients with cyanosis of unknown cause5. Angiography is unable to detect fistulae below 0.5 mm diameter. In those cases, diagnostic may be obtain with a combination of tests such as contrasted echocardiography, CT-scan, MRI, scintigraphy with radioisotopes and lung biopsy. CT-scan identifies 98% of PAVF while angiography identifies 60%. In spite of this, angiography is important to determine fistulae site and size, and when surgical treatment is a possibility6.

Non-treated lesions are associated to mortality in 11% and morbidity in 26% of cases. Relative indication for treatment includes asymptomatic patients with lesions less than 1 cm with minimum or no shunt. Absolute indication is reserved to symptomatic patients with multiple large or growing fistulae, with systemic circulation repercussion, presence of major hypoxemia in room air, complicated PAVF and Rendu-Osler-Weber syndrome patients5. Treatment may be surgical or by embolization. Embolization may be achieved with coil (type of spring filling the fistula) or cuff5. Central and extensive PAVF, with short neck, active bleeding by intrapleural rupture and repeated failed embolizations are best treated with surgery1. Surgical treatment includes lobectomy, fistulae resection or ligation5. Some authors, however, refer that embolization is a safe and effective treatment for large central PAVF with or without short pedicle7.

In our case, after MRI to elucidate changes observed at chest X-rays, and pulmonary arteriography for surgical planning, upper left lobectomy was the choice. Indicated anesthetic technique was intravenous general anesthesia with propofol, atracurium and fentanyl associated to spinal anesthesia with morphine and hyperbaric bupivacaine. Blockade advantages in this case were intraoperative hemodynamic stability and anesthesia, decreasing the need for opioids during the procedure, postoperative analgesia with minor side effects and allowing early tracheal extubation. Spinal blockades with morphine provide analgesia, help respiratory physical therapy with better pulmonary expansion and decreased atelectasis areas. In pediatric patients, even with spinal anesthesia at T3 to T5 level, local anesthetics, associated or not to opioids, determine minor heart rate and blood pressure changes, which are easily corrected with fluids or alpha-agonists for hypotension, and anti-cholinergic drugs for bradycardia. This is due to poor sympathetic lower limbs innervation or sympathetic nervous system immaturity in children and neonates8.

It is known that monopulmonary ventilation is highly desirable during this type of surgery because lung deflation improves visualization of chest contents and may decrease the risk for pulmonary injuries induced by retractors. In children, a conventional tracheal tube may be used for selectivity. In intubating the left bronchium, there must be an 180º tube rotation and head should be turned to the right. Tracheal tube should advance in the bronchium until disappearance of the sound in the side to be operated on. A bronchoscope may be used to check intubation or to guide positioning. This is a simple technique not needing special equipment in addition to the bronchoscope, since in Brazil we do not have pediatric double-lumen tracheal tubes. Problems which may be seen with the use of conventional tracheal tubes for selective intubation are that, if small cuffless tubes are used, there might be difficult to obtain adequate intubated bronchium seal9. In our case, monopulmonary ventilation was achieved without intercurrences with selective intubation to the right to help surgical exposure.

Intraoperative blood gas analysis evolution is shown in table I.

Initial blood gas revealed hypoxemia with low PaO2. There has been hypoxemia worsening during monopulmonary ventilation which as only improved with bipulmonary ventilation. Blood gas were normal at surgery completion and the child was extubated one hour later. As the evolution was progressive decrease in peripheral saturation due to the increase of the larger fistula, embolization was performed with coil.

The treatment of the first fistula has allowed the appearance of other fistulae and others may appear in the lungs. Extrapulmonary arterial venous fistulae may determine the diagnosis of Rendu-Osler-Weber syndrome.



01. Pick A, Deschamps C, Stanson AW - Pulmonary arteriovenous fistula: presentation, diagnosis, and treatment. World J Surg, 1999;23:1118-1122.        [ Links ]

02. Amaral FTV, Felix PR, Granzotti JA et al - Fístula artério-venosa pulmonar maciça. Causa rara, potencialmente curável de hipóxia neonatal. Arq Bras Cardiol, 1996;66:353-355.        [ Links ]

03. Marchesani F, Cecarini L, Pela R et al - Pulmonary arteriovenous fistula in a patient with Rendu-Osler-Weber syndrome. Respiration, 1997;64:367-370.        [ Links ]

04. Berg JN, Guttmacher AE, Marchuk DA et al - Clinical heterogeneity in hereditary haemorrhagic telangiectasia: are pulmonary arteriovenous malformations more common in families linked to endoglin? J Med Genet, 1996;33:256-257.        [ Links ]

05. Hirata K, Yamagishi H, Akioka K et al - A case of diffuse pulmonary arteriovenous fistula. Jpn Circ J, 1999;63:499-501.        [ Links ]

06. Hamada H, Terai M, Okajima Y et al - Angiographical and computed tomographic findings in diffuse pulmonary arteriovenous fistulas. Int J Card, 1997;59:203-205.        [ Links ]

07. Clark JA, Pugash RA, Faughnan ME et al - Multidisciplinary team interested in the treatment of pulmonary arteriovenous fistulas or malformations (PAVFs). World J Surg, 2001;25: 254-255.        [ Links ]

08. Caires JL - Raquianestesia associada à anestesia geral em cirurgia cardíaca. CEDAR, 2001;15:16-18.        [ Links ]

09. Hammer GB - Pediatric thoracic anesthesia. Anesth Analg, 2001;92:1449-1464.        [ Links ]



Correspondence to
Dra. Maria José Carvalho Carmona
Address: Rua Rodésia, 161/82 Vila Madalena
ZIP: 05435-020 City: São Paulo, Brazil

Submitted for publication July 25, 2003
Accepted for publication December 15, 2003



* Received from Serviço de Anestesiologia e Divisão de Cirurgia do Instituto do Coração do Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo (HCFMUSP), SP

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License