Article Text

PDF

Isolated pulmonary arteriovenous fistula without Rendu-Osler-Weber disease as a cause of cryptogenic stroke
  1. K Kimura,
  2. K Minematsu,
  3. M Nakajima
  1. Cerebrovascular Division, Department of Medicine, National Cardiovascular Center, Fujishirodai, Suita, Osaka, Japan
  1. Correspondence to:
 Dr K Kimura
 Cerebrovascular Division, Department of Medicine, National Cardiovascular Center, 5-7-1 Fujishirodai, Suita, Osaka 565-8565, Japan; kimurakhsp.ncvc.go.jp

Abstract

There has been uncertainty as to whether a right to left shunt through an isolated pulmonary arteriovenous fistula (P-AVF) without Rendu-Osler-Weber (ROW) disease can cause paradoxical brain embolism. A population of 747 acute ischaemic stroke patients was examined to determine the frequency and clinical characteristics of those patients who had an isolated P-AVF. The presence of a P-AVF was determined as follows. On patients with a stroke of undetermined cause, both transoesophageal echocardiography and transcranial Doppler with saline contrast medium was performed to detect a right to left shunt. If a P-AVF was then suspected, selective pulmonary angiography and enhanced chest CT was performed to confirm the presence of the P-AVF. Four patients (0.5%) were diagnosed as having a stroke associated with an isolated P-AVF. All the patients were middle-aged women (mean age 61 years). In all these patients, the P-AVF could not have been suspected on physical findings or chest x ray. The P-AVF was always single and located in the lower lobe. All the patients had asymptomatic deep venous thrombosis, and three patients developed pulmonary embolism. As d-dimer and thrombin–antithrombin complex were elevated in all patients, this indicated an activation of both fibrinolytic and thrombin activity. Our results show that an isolated P-AVF without ROW disease can cause paradoxical brain embolism. Thus, the existence of an isolated P-AVF as a right to left shunt in patients with a stroke of unknown origin should not be overlooked, even if a P-AVF is not suggested by the initial physical findings or chest x ray.

  • Rendu-Osler-Weber disease
  • cryptogenic stroke
  • pulmonary arteriovenous fistula
  • DVT, deep venous thrombosis
  • MES, microembolic signals
  • P-AVF, pulmonary arteriovenous fistula
  • PFO, patent foramen ovale
  • ROW, Rendu-Osler-Weber
  • TCD, transcranial Doppler
  • TOE, transoesophageal echocardiography

Statistics from Altmetric.com

Rendu-Osler-Weber (ROW) disease is characterised by multiple dermal, mucosal, and visceral telangiectasia that are associated with recurrent bleeding, and by a pulmonary arteriovenous fistula (P-AVF) in 15% of patients.1 The right to left shunt caused by P-AVF with ROW disease can cause paradoxical brain embolism.1,2 To the best of our knowledge, however, only four cases of paradoxical brain embolism associated with an isolated P-AVF without ROW disease, including our one previous case, have been reported.3–6 Therefore, the question remains as to whether an isolated P-AVF is associated with ischaemic stroke, and, in particular, with paradoxical brain embolism. The aims of this study were to investigate the frequency of brain infarction associated with an isolated P-AVF, to evaluate the salient clinical characteristics of this condition, and to elucidate a mechanism for the development of ischaemic stroke in these patients.

SUBJECTS AND METHODS

We reviewed the records of 747 consecutive ischaemic stroke patients who were admitted to our division within 7 days of stroke onset between August 1998 and December 2002. We identified those patients with a brain infarction that was associated with an isolated P-AVF without ROW disease. The diagnosis of ROW disease was made clinically based on the “classic triad” of telangiectasia, recurrent epistaxis, and a family history of the disorder. The presence of a P-AVF was determined as follows. When we had a patient with a stroke of undetermined cause, we always performed both a transoesophageal echocardiography (TOE) with saline contrast medium and a transcranial Doppler (TCD) with saline contrast medium so as to detect a right to left shunt, such as a patent foramen ovale (PFO) or a P-AVF. A patient with a stroke of undetermined cause was defined as a patient who did not have a lacunar stroke; did not have more than a 50% stenosis in the cerebral artery irrigating the affected lesions; and had no potential cardiac sources of emboli (such as atrial fibrillation, acute myocardial infarction, old myocardial infarction with intraventricular thrombus, mitral valve disease, prosthetic valve, implantation of a pacemaker, or a dilated cardiomyopathy). If a P-AVF was suspected based on the findings of the TCD study with saline contrast medium that detected microembolic signals (MES) through the middle cerebral artery or the basilar artery, and the TEE did not demonstrate a PFO,7 we then always performed selective pulmonary angiography and enhanced chest CT so as to confirm the presence of the P-AVF. Pulmonary embolism was diagnosed by lung perfusion scintigraphy, and deep venous thrombosis (DVT) was diagnosed by venography and/or ultrasonography.

RESULTS

ROW disease was not observed in the 747 patients studied. However, seven patients were suspected as having a P-AVF by TCD and TEE studies. Pulmonary angiography and chest CT studies could not confirm a P-AVF in three of these patients. Therefore, four patients (0.5%), including our previously reported patient (case 1),6 were diagnosed as having an embolic stroke associated with an isolated P-AVF. TCD studies in all these patients showed that MES were detected during normal breathing without having to perform a Valsalva manoeuvre or a cough. The clinical characteristics of all four patients are shown in the table. None of the patients had clinical evidence of hypoxia, such as cyanosis, dyspnea, and erythrocytosis. We could not auscultate vascular sounds in the lung fields, and the chest x rays did not demonstrate any abnormality, such as a nodular density in the bilateral lung lobes, which would have led us to suspect P-AVF. The P-AVF found on lung CT and pulmonary angiography was always single and located in the lower lobe. The mean size of the P-AVF on CT was 4.5 mm. All patients had pulmonary embolism, and three had asymptomatic deep venous thrombosis. d-dimer and thrombin–antithrombin complex were elevated in all patients, indicating the activation of both fibrinolytic and thrombin activity.

Table 1

Clinical characteristics of four patients with P-AVF without Rendu-Osler-Weber disease

The P-AVFs were occluded with catheter embolisation within 2 months of stroke onset. Our follow up was conducted for an average of 32 months (range 13–57 months), and no recurrent ischaemic strokes were documented, but one patient (case 4) had a thalamic haemorrhage 8 months after catheter embolisation of the P-AVF. This patient had been treated with warfarin and her international normalised ratio was 2.2 at the onset of the brain haemorrhage.

Fig 1 shows the findings of case 2: a pulmonary angiography with a P-AVF; a venogram of the lower limbs showing a DVT; and lung perfusion scintigraphy showing pulmonary embolism.

Figure 1

The pulmonary angiography (A, B), venogram of the lower limbs (C), and the ventilation–perfusion lung scintigraphy (D, E) of case 2. (A) Selective pulmonary angiography 30 days after stroke onset shows a P-AVF (arrow) in the right lower lobe. (B) After embolisation therapy with a metal coil, the feeding vessels to the P-AVF are completely occluded. (C) Venogram of the left lower limbs 4 days after stroke onset shows the abnormal collateral flow (arrow) in the leg veins, indicating a deep venous thrombus. (D) Lung perfusion scintigram 4 days after stroke onset shows the defect (arrow) in the left upper lung, indicating a pulmonary embolism. (E) The follow-up study taken 24 days after stroke onset shows no perfusion defect in the left upper lung.

DISCUSSION

We have presented four acute ischaemic stroke patients with P-AVF not associated with ROW disease. None of the patients had cardiac or arterial sources for the emboli. Furthermore, all the patients had DVTs and three patients developed pulmonary embolism. Therefore, we diagnosed all these patients as having had paradoxical brain embolism through a P-AVF. We conclude that an isolated P-AVF can cause paradoxical brain embolism.

It is of interest to note that the size of the P-AVFs differed between previously reported cases and our patients. The previously reported cases, due to the large size of the P-AVFs,3–5 had physical findings, such as auscultate vascular sounds, or abnormalities on chest x ray. In contrast, in all of our patients the P-AVFs could not be suspected on physical findings or chest x ray, because of their small size. Therefore, the presence of P-AVF in stroke patients who have normal physical findings and no chest x ray abnormalities cannot be excluded.

A TCD with saline contrast medium performed on all patients was useful in identifying the presence of a P-AVF during the acute phase of the stroke. As a persistent right to left shunt occurs in a P-AVF, micro air bubbles can be detected in the cerebral arteries using TCD with saline contrast medium during normal breathing without the need to use provocative methods, such as a Valsalva manoeuvre or a cough.7 Therefore, in patients who have had an embolic stroke of undetermined cause and cannot perform a Valsalva manoeuvrer or a cough because of aphasia or a disturbance of consciousness, TCD with saline contrast medium can help detect a P-AVF in the acute phase of stroke.

Catheter embolisation is a safe and effective treatment for P-AVF.8 All four patients had a history of ischaemic stroke or TIA prior to the present stroke. After catheter embolisation of their P-AVF, none of these patients had a recurrent ischaemic stroke. Thus, catheter embolisation of a P-AVF appears to be an effective method of preventing recurrent ischaemic stroke in such patients.

In conclusion, an isolated P-AVF without ROW disease can cause cryptogenic stroke. Thus, one should not overlook an isolated P-AVF as a right to left shunt in patients with a stroke of unknown origin, even when physical findings or chest x ray findings are not suggestive of a P-AVF.

Acknowledgments

This study was supported in part by Research Grants for Cardiovascular Disease (12A-4, 14C-1) from the Ministry of Health, Labor and Welfare of Japan.

REFERENCES

View Abstract

Footnotes

  • Competing interest: none declared

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.