| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
Dr. Gossage is Director, Pulmonary Vascular Diseases, and Director, Multidisciplinary ICU, Section of Pulmonary and Critical Care Medicine, Department of Medicine, Medical College of Georgia.
Correspondence to: James R. Gossage, MD, FCCP, Director, Pulmonary Vascular Diseases, Director, Multidisciplinary ICU, Medical College of Georgia, Section of Pulmonary and Critical Care Med, BBR-5513, 1120 Fifteenth St, Augusta, GA 30912-3135; e-mail: jgossage{at}mail.mcg.edu
Pulmonary arteriovenous malformations (PAVMs) represent a direct communication between one or more pulmonary arteries and one or more pulmonary veins. Common anatomic forms include single sacs ranging from 1 to > 10 cm in diameter, macroscopic tangles, and microscopic telangiectases. The main complications of PAVMs are believed to relate to right-to-left shunting of blood through the PAVM and include stroke and brain abscess, which together occur in up to one half of patients if untreated.1 Other potentially serious complications include refractory hypoxemia, hemoptysis, and hemothorax. Fortunately, nearly all of these complications can be prevented by embolization of the feeding arteries (transcatheter embolotherapy [TCET]). PAVMs are associated with underlying hereditary hemorrhagic telangiectasia (HHT) in up to 90% of cases. Conversely, at least 30% of patients with HHT will also have PAVMs. In order to prevent the complications noted above, various screening strategies for PAVMs have been proposed in patients with known HHT or their undiagnosed relatives.1 2 3 4 The main screening tests are reviewed in the following paragraphs.
Contrast transthoracic echocardiography (TTE) is performed by IV injection of echocardiographic contrast while visualizing the atria. In the case of PAVM, contrast appears in the left atrium 2 to 5 s after it is seen in the right atrium. Nanthakumar and colleagues5 screened 106 HHT patients with contrast TTE, oxygen shunt test, and chest radiography. All patients with one or more positive screening test results underwent standard pulmonary angiography. Contrast TTE findings were positive in 33 of 35 patients with PAVMs (94% sensitivity), and was the only positive test result in 11 patients with PAVMs; it is unknown what proportion of these PAVMs were amenable to TCET. In a study by Kjeldsen and colleagues,4 48% of HHT patients with a positive contrast TTE result were found to have PAVMs that were amenable to TCET. Overall, contrast TTE has the advantages of exquisite sensitivity and good availability, but is fairly expensive and lacks specificity for treatable PAVMs.
Right-to-left shunt can be calculated by the oxygen shunt test (OST) by measuring a single arterial blood gas after a patient breathes 100% oxygen for 15 to 20 min. In a small screening study, Haitjema and colleagues2 found the OST to have a sensitivity of 88% for PAVM. Kjeldsen and colleagues4 evaluated various screening procedures in 25 patients with HHT who had a positive contrast TTE result; 15 patients had PAVMs by angiography. A PaO2 < 500 mm Hg during an OST had a sensitivity of 100% and a specificity of 40%. Since they evaluated only patients with a positive contrast TTE result, the true sensitivity and specificity in a screening population are unknown. However, in a review of the literature, a shunt of > 5% by OST was seen in 97.5% of patients who had a PAVM large enough for TCET.1 The OST has the advantages of low expense and universal availability, but may overestimate shunt if technique is not adhered to.1 Some investigators have considered the OST to be invasive, but it is really only slightly more invasive than the other two screening tests that require an IV injection. Measurement of PaO2 while breathing room air or use of pulse oximetry has inadequate sensitivity for screening.4
Radionuclide perfusion lung scanning (PLS) can be used to estimate shunt after IV injection of radionuclide labeled macroaggregated albumin by measuring differential perfusion to the lungs vs either brain or kidneys. It shows excellent agreement with the percentage of shunt determined by OST.6 Thompson and colleagues3 evaluated 66 patients after TCET for PAVMs. Forty patients had small residual PAVMs with feeding arteries < 3 mm, while 26 patients had no residual PAVMs. In patients with residual PAVMs, PLS demonstrated a mean shunt of 9.3%. A shunt > 3.5% was 87% sensitive and 61% specific for residual PAVM. This method would almost certainly be even more sensitive for detecting larger PAVMs prior to TCET. Radionuclide perfusion lung scanning and contrast TTE have the advantage of being fairly specific for anatomic shunt (in contrast to OST, which does not discriminate between anatomic and physiologic shunt). Because it is quantitative, PLS could theoretically show improved specificity over contrast TTE for treatable PAVM; however, PLS is not widely available and is also expensive.
In this issue of CHEST (see page 351), Lee and colleagues (see page 342) report on the use of contrast TTE in the follow-up of 39 patients after TCET. The authors screened all referrals to their HHT clinic for PAVMs with chest radiography, OST, and contrast TTE, and subsequently performed pulmonary angiography in all patients with one or more positive screening test result. Contrast TTE was 100% sensitive in detecting PAVMs vs the OST, which was only 59% sensitive. Twenty-nine patients underwent contrast TTE both before and after TCET. Although all PAVMs with feeding arteries 
3 mm were successfully occluded during TCET—an average of five PAVMs per patient—52% of patients had residual PAVMs with feeding arteries < 3 mm. Following TCET, contrast TTE remained positive in 90% of patients while the OST result was positive in 27%. The results were not significantly different for those with and those without residual PAVM. The authors nicely reviewed and addressed potential limitations of their study, the most important being that only 29 of the 43 patients undergoing TCET had pre-TCET and post-TCET echocardiograms. They concluded that contrast TTE findings remain positive in most patients after TCET, suggesting the presence of at least microscopic PAVMs even in those without residual PAVMs on angiography.
This is a well-done study that makes an important contribution to the literature on screening of HHT patients and follow-up strategy for patients with known PAVMs. It confirms the high sensitivity of contrast TTE in detecting PAVM. It also confirms the widely-held suspicion that patients often demonstrate persistent shunt after successful TCET. The only criticism I have of their study relates to the cutoff used for the OST, which was based on optimal accuracy from a receiver operator characteristic curve. They chose an alveolar-arterial gradient of 175, which equates to a PaO2 of approximately 488 mm Hg and a shunt of approximately 9.5%.1 Since numerous studies have shown that patients with treatable PAVMs may have a shunt at least as low as 5%, their cutoff was predestined to "stack the deck" against the OST. Although a grading system has not yet been proposed for estimating shunt during contrast TTE, it seems likely that lower grades would be associated with less shunt and less clinical significance, as is the case for the OST. If they had used a PaO2 of 575 mm Hg (which equates to a shunt of approximately 5%) for their cutoff, they may have picked up more PAVMs during the screening process, which may have affected the sensitivity of contrast TTE.
There are several problems with interpretation of the current literature on screening for PAVM. First, all screening studies have lacked a control group without suspected HHT or PAVMs, and therefore true specificity cannot be determined. Second, no study has subjected all screened patients to standard pulmonary angiography; therefore, the true sensitivity for treatable PAVM cannot be determined.
Based on current data, I believe that contrast TTE is the best initial screening test due to its excellent sensitivity and availability. If the result is negative, the likelihood of significant PAVM is low. If suspicion for PAVM remains high despite a negative contrast TTE result, additional screening tests could be performed. If the contrast TTE result is positive, the patient is very likely to have at least microscopic PAVMs.4 Current contrast TTE techniques are not quantitative, however, and therefore do not predict the amenability of PAVMs to TCET. I can envision three general approaches to patients with a positive contrast TTE result. The most aggressive approach would subject all such patients to standard pulmonary angiography. A middle approach would subject all patients to contrast CT angiography followed by standard pulmonary angiography when the CT finding is positive for significant PAVMs. A conservative approach might involve additional testing to determine the likelihood that the PAVMs are amenable to TCET. This testing could include chest radiography, OST, PLS, CT angiography, or MRI; patients with a second positive test result would then undergo standard pulmonary angiography. Which of these approaches has the optimal accuracy, clinical benefit, and cost-effectiveness cannot be determined from current data. Alternatively, if cost is a concern, initial screening could be done with the OST, followed by either contrast TTE or PLS.1
What about follow-up after TCET? Lee and colleagues recommended life-long follow-up and antibiotic prophylaxis in patients with a persistently positive contrast TTE findings after TCET. I would go one step further. Since at least 90% of patients with HHT and PAVMs will have a persistently positive contrast TTE findings,7 and since the natural history of those with a negative contrast TTE result is unknown, I would actually recommend long-term follow-up and antibiotic prophylaxis in all patients with HHT and PAVMs. The other goal of follow-up is to detect development of PAVMs that are amenable to TCET. I recommend the OST for this goal since it is fairly sensitive to treatable PAVMs,1 but not so sensitive that it picks up all PAVMs.
References
This article has been cited by other articles:
|
|
 |
|
  N. W. Morrell Screening for Pulmonary Arteriovenous Malformations Am. J. Respir. Crit. Care Med., May 1, 2004; 169(9): 978 - 979. [Full Text] [PDF]
|
|
|
|
|
|
V. Cottin, H. Plauchu, J.-Y. Bayle, M. Barthelet, D. Revel, and J.-F. Cordier Pulmonary Arteriovenous Malformations in Patients with Hereditary Hemorrhagic Telangiectasia Am. J. Respir. Crit. Care Med., May 1, 2004; 169(9): 994 - 1000. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
