Telangiectasia, hereditary hemorrhagic

Pulmonary arteriovenous malformations (PAVMs) associated with hereditary hemorrhagic telangiectasia may cause severe cerebral complications that may be prevented by embolization therapy. We retrospectively compared the diagnostic value of noninvasive tests for the screening of treatable (amenable to embolization) PAVMs in a series of 105 patients, using chest computerized tomography (CT) and/or pulmonary angiography as a "gold standard." Patients had assessment of dyspnea, chest radiograph, alveolar-arterial PO^sub 2^ gradient under 100% oxygen (AaPO^sub 2^), contrast echocardiography, and radionuclide perfusion lung scanning. Contrast echocardiography in the supine position was the most sensitive test (93%). The sensitivity of self-reported dyspnea (59%), chest radiograph alone (70%), measurement of AaPO^sub 2^ by the 100% oxygen method (62%), or radionuclide lung scanning (71%), was not suitable for efficient screening. A 100% sensitivity and negative predictive value could be obtained when combining anteroposterior chest radiograph and contrast echocardiography. Our data support a screening algorithm based on the combined use of contrast echocardiography and anteroposterior chest radiograph, followed by chest CT if either test is positive. An alternative is to screen directly by chest CT. However, this algorithm may obviate the need for chest CT in patients without PAVM, who represent a majority of patients with hereditary hemorrhagic telangiectasia.

Keywords: contrast echocardiography; hereditary hemorrhagic telangiectasia; Osler-Rendu-Weber disease; pulmonary arteriovenous malformations; right-to-left shunting

Hereditary hemorrhagic telangiectasia (HHT), or Osler-Rendu-Weber disease, is an autosomal dominant disorder characterized by recurrent epistaxes, mucocutaneous telangiectasias, and visceral involvement (1, 2) including arteriovenous communications that may develop in virtually any organ, especially in the lung. The prevalence of HHT may exceed 1 in 10,000 in some regions (3). The prevalence of pulmonary arleriovenous malformations (PAVMs) has been estimated as between 15 and 33% in previous studies (3, 4). PAVMs may enlarge over time, especially during pregnancy (5). PAVMs may cause hypoxemia and dyspnea due to right-to-left shunting, but remain frequently undiagnosed. However, severe complications may occur, such as massive hemoptysis or hemothorax, and especially neurologic complications including transient ischemic attack, cerebral stroke, and cerebral abscess due to the right-to-left shunting that facilitates the passage of septic emboli into the cerebral circulation. Such severe complications may be the presenting manifestation leading to diagnosis of the PAVM and even of HHT itself (2). Cerebral stroke and abscess have been reported in up to 30 and 5-9% of patients with HHT and PAVM, respectively (6, 7). The risk of cerebral complications is usually considered significant when the feeding artery of the PAVM exceeds 3 mm in diameter, and may increase in patients with multiple PAVMs (8). More than two-thirds of neurologic manifestations of HHT are related to PAVM, with the remaining one-third due to cerebral or spinal arteriovenous malformations that may cause subarachnoid hemorrhage or seizure (9). Treatment of PAVMs is thus justified, even if PAVMs are asymptomatic, when the diameter of the feeding vessel is more than 3 mm. Transcatheter occlusion of PAVMs, especially using steel coils (embolization therapy), usually obviates the need for thoracotomy, is well tolerated, significantly decreases right-to-left shunting (10, 11), and may reduce the risk of cerebral complications (3).

Given the high risk of cerebral abscesses and ischemic complications of PAVM that can be reduced by an effective treatment, screening of patients with HHT for asymptomatic PAVM has been proposed (7), especially for women of childbearing age (12). However, no routine screening algorithm has been established. Tests commonly used for screening include chest radiograph, arterial oxygen measurement on 100% oxygen, contrast echocardiography, and radionuclide lung scanning, but their accuracy has not been compared with reference to a "gold standard" (13). In particular, the diagnostic value of contrast echocardiography has not been systematically compared with chest CT and/or pulmonary angiography (14, 15). Chest CT, which is at least as sensitive and specific as pulmonary angiography for the diagnosis of PAVM (16), may now be considered the gold standard for the diagnosis of PAVM. Pulmonary angiography is thus indicated only when treatment by embolization therapy is considered.

We conducted a retrospective study to compare the diagnostic value of noninvasive tests for the screening of treatable PAVM (i.e., PAVMs that are amenable to embolization therapy) in patients with HHT, to establish an evidence-based algorithm for screening and diagnosis of PAVM in HHT. Some of the results of this study have been previously reported in the form of an abstract (17).

METHODS

Patients

We retrospectively studied a consecutive series of patients referred between January 1, 1982 and May 1, 2003 with a diagnosis of HHT and who underwent screening or diagnostic tests for PAVM. The diagnosis of HHT was based on published criteria (1). Patients had chest radiograph, pulmonary function tests, alveolar-arterial PO^sub 2^ gradient (AaPO^sub 2^), contrast echocardiography, radionuclide perfusion lung scanning, and CT of the chest, done within 1 week. Pulmonary angiography was performed as part of the therapeutic procedure in patients with PAVM on CT and in whom embolization therapy was indicated. The diagnostic tests were performed without the knowledge of the result of the reference standard test. All patients (or parents for patients younger than 18 years) gave oral consent for the tests after detailed information was provided.

Diagnostic Tests

Pulmonary function tests were performed as recommended (18). For AaPo2 measurement, PA^sub O^sub 2^^ was measured after the patient had been breathing 100% O2 for at least 15 minutes, with a deep inspiration every minute. Actual PI^sub O^sub 2^^ was measured to estimate PA^sub O^sub 2^^ with PA^sub O^sub 2^^ (kPa) = PIO^sub 2^ - 6.27 - (Pa^sub CO^sub 2^^/0.8), and AaPO^sub 2^ = PA^sub O^sub 2^^ - Pa^sub O^sub 2^^. In our laboratory, normal values of AaPO^sub 2^ with 95% confidence are less than 18.6 and 24.5 kPa in the supine and upright positions, respectively (19).

Spiral volumetric CT scan of the chest was performed with the single breath-holding technique (16). Intravenous administration of contrast medium was used only for evaluation of large PAVMs (n = 29), or for the differential diagnosis of nodules (n = 24). The diagnosis of PAVM was based on the presence of a mass or nodule fed by an enlarged artery, with characteristic enhancement after contrast medium injection when performed (16).

Transthoracic contrast echocardiography was performed by injecting 4.5 ml of agitated modified fluid gelatin solution with 0.5 ml of room air into a peripheral vein while simultaneously imaging the atria by two-dimensional echocardiography. The test was considered positive for pulmonary right-to-left shunting when contrast was visualized in the left atrium after a delay of at least four cardiac cycles.

Radionuclide perfusion lung scanning was performed in the supine position after bolus injection of ^sup 99m^Tc-labeled macroaggregates of albumin into an antecubital vein. The distribution of the macroaggregates was measured over the lungs, kidneys, and cerebral areas.

Digital subtraction pulmonary angiography was performed with a 7F catheter introduced into each pulmonary artery under radioscopic guidance, and injection of 20-25 ml of iodinated contrast medium alternatively in each pulmonary artery.

Because chest CT has been shown to be at least as sensitive and specific as pulmonary angiography for the diagnosis of PAVM (16), the gold standard for the diagnosis of PAVM in the present study was defined as a positive CT scan of the chest and/or positive pulmonary angiography, with either test demonstrating the presence of at least one PAVM that was amenable to embolization therapy (i.e., PAVM with a feeding vessel of at least 2-3 mm). Only patients for whom CT scan of the chest and/or pulmonary angiography were available were included in this study.

RESULTS

Study Population

One hundred and sixteen consecutive patients underwent screening or diagnostic tests for PAVM during the study period. Eleven patients were excluded from the study because neither chest CT nor pulmonary angiography had been performed (n = 10), or because of incomplete medical chart (n = 1). At least two diagnostic criteria for HHT were present in all patients (Table 1). Thus, the population available for analysis consisted of 105 patients.

In 85% of the patients, the diagnosis of HHT was made before the referral for screening of PAVM, including two-thirds of asymptomatic patients with no known radiographie abnormalities, and 21% in whom the diagnostic tests were indicated because of a respiratory symptom (9%), an opacity on systematic chest radiograph (7%), or neurologic manifestations (5%). In 15% of the cases, the patients were not aware of a diagnosis of HHT until they were referred to our institution, where the diagnosis of HHT was established.

Of 70 women, 24 had been pregnant, with a mean number of pregnancies of 1.61 ± 1.3, and 9 were nulliparous; the information was not recorded for 36. One patient was pregnant at the time the diagnostic tests were performed. In only two cases were the screening tests clearly motivated by a wish to have these done before having a child.

The main clinical manifestations are shown in Table 1. Dyspnea was by far the most frequent respiratory symptom, present in half the patients, with New York Heart Association classes as follows: Class I, 10 patients; Class II, 36 patients; Class III, 5 patients; Class IV, none. Thirty-nine percent of the patients were current or ex-smokers.

Diagnosis of PAVM Amenable to Embolization Therapy

According to the inclusion criteria, all patients included in the study had a CT scan of the chest (n = 99) performed as a routine diagnostic test of PAVM related to HHT and/or pulmonary angiography (n = 47). The CT scan showed at least one PAVM in 45 patients, including 41 in whom pulmonary angiography (performed as a therapeutic test) confirmed the diagnosis of PAVM, and 4 patients in whom pulmonary angiography was not performed. In addition, a PAVM was diagnosed by pulmonary angiography in six patients in whom CT scan was not performed. Results of the CT scan and of pulmonary angiography were consistent in all 41 patients in whom both tests were performed. No PAVM amenable to embolization therapy was found on CT scan of 54 patients in whom pulmonary angiography was not performed. Of note, nonspecific micronodular opacities (less than 3 mm, with no feeding vessel visible) were present on chest CT scan of five patients, potentially representing small PAVMs not amenable to embolization therapy. Because the objective of the present study was to detect treatable PAVM, those patients were considered "PAVM-negative" for the analysis. Thus, PAVMs amenable to embolization therapy were diagnosed by CT scan and/or pulmonary angiography in 51 of 105 patients (48.6%).

Diagnostic Value of Noninvasive Tests

The diagnostic value of noninvasive tests is presented in Table 2. The accuracy (i.e., the proportion of patients correctly classified) for the diagnosis of PAVM was 86% for positive anteroposterior chest radiography and positive radionuclicle lung scanning, 83% for an abnormal AaPO^sub 2^ gradient, and 67% for echocardiography in the supine position. Dyspnea on exertion had an accuracy of only 60%.

Among noninvasive tests, contrast echocardiography in the supine position had the best diagnostic values required for a screening test, with a sensitivity of 92%, a predictive value for a negative test of 97% (in our group of patients with a high prevalence of PAVMs: 48.6%), and a negative likelihood ratio of 0.17. Performing contrast echocardiography in both the supine and upright positions added little to its sensitivity (no patient with PAVM had positive echocardiography in the upright position and negative echocardiography in the supine position). Assessment of dyspnea, chest radiograph, radionuclide lung scanning, and detection of right-to-left shunting by the 100% oxygen method (AaPO^sub 2^) were less efficient as screening tests, with a sensitivity lower than 75% and a negative predictive value lower than or equal to 80%. Contrast echocardiography was falsely negative in both the upright and supine positions in a single patient, in whom radionuclide lung scanning showed positive activity on both renal and cerebral areas, and chest radiograph and CT scan showed a single PAVM, which was confirmed by pulmonary angiography; AaPO^sub 2^ on 100% oxygen was normal in this patient.

The assessment of right-to-left shunting by the isotopic method was highly specific, with a specificity of 98%, a predictive value of a positive test of 97%, and a positive likelihood ratio of 24 (Table 2).

In this group of patients with HHT, the presence of suggestive opacities on chest radiograph was also highly specific for PAVM, with a specificity of 98% and a positive predictive test of 97%, whereas sensitivity was lower (70%). Performing the chest radiograph with both anteroposterior and lateral views did not add to its diagnostic value (data not shown).

Contrast echocardiography was considered falsely positive in 25 patients without PAVM on reference CT scan, including 2 with micronodular opacities (2 and 3 micronodules, respectively) that might correspond to minute PAVM not amenable to embolizalion therapy, and only 1 with patent foramen ovale and ancurysm of the interatrial septum. Chest radiograph and AaPO^sub 2^ were normal in all these patients, and radionuclide lung scanning was normal in all but one patient. Intracardiac shunting was also found in three patients with PAVM, showing that intracardiac and intrapulmonary right-lo-left shunting may be associated.

Chest radiograph was considered falsely positive in a single patient (chest radiograph and CT scan were normal 6 months later). Radionuclide lung scanning was considered falsely positive in a single patient with normal CT scan. Measurement of AaPo2 gradient on 100% oxygen also gave a single false positive result when using normal values from our laboratory (AaPO^sub 2^ was 18.6 kPa; normal,

It is noteworthy that the rate of positive tests was low in the group of 54 patients with a normal CT scan of the chest and in whom a pulmonary angiogram was not performed [i.e., contrast echocardiography, radionuclide lung scanning, and AaPo2 with a single false positive case each]; on the basis of this observation, it is highly unlikely that hemodynamically significant PAVM may have been missed by the CT scan of the chest.

Diagnostic Value of AaPO^sub 2^ (100% O2)

A receiver-operating-characteristic curve analysis was performed for the quantitative measurement of AaPO^sub 2^ under 100% O2. This analysis shows the fraction of true positive results (sensitivity) and false positive results (1 - specificity) for various cutoff levels of AaPO^sub 2^. The threshold that gave the maximal accuracy for AaPO^sub 2^ (100% O2) for the supine position was 9.1 kPa. At this threshold the sensitivity was 80% and the specificity was 92% (Figure 1). The threshold for AaPO^sub 2^ (100% O2) for the upright position was 9.7 kPa, with a sensitivity of 85% and a specificity of 86%. The calculated area under the curve was 0.90 (95% confidence interval, 0.84-0.96) for the supine position and 0.92 (95% confidence interval, 0.87-0.98) for the upright position, where a value of 0.5 is no better than expected by chance, and a value of 1.0 reflects a perfect indicator. Of note, similar results were obtained when AaPO^sub 2^ was estimated on room air, a simpler method to estimate right-to-left shunting (data not shown).

Low thresholds of 5.88 and 5.17 kPa in the supine and upright positions, respectively, were required to obtain a sensitivity of 90%, with a lower specificity (71 and 65%, respectively). When normal values (95% confidence) from our laboratory were used (18.6 and 24.5 kPa in the supine and upright positions, respectively), corresponding to a right-to-left shunt of 5% of cardiac output, excellent results were obtained for specificity (98%), predictive value of a positive test (97%), and positive likelihood ratio (20) (Table 2).

Thus, the measurement of AaPO^sub 2^ (100% O2) was less sensitive than contrast echocardiography and equally specific to radionuclide lung scan and chest radiograph, indicating that this test may not be the best suited for screening patients with HHT. Normal values from our laboratory (rather than arbitrary lower thresholds) were used thereafter in the study (Tables 2 and 3).

Diagnostic Value of Combined Noninvasive Tests

As shown in Table 3, the combination of anteroposterior chest radiograph and contrast echocardiography in the supine position had the best diagnostic values required for a screening test, with a sensitivity of 100%, a predictive value of a negative result of 100%, and a likelihood ratio of a negative test result of 0.04. Hence, none of the patients in our series had PAVM on CT scan of the chest with both negative chest radiograph and contrast echocardiography. The combination of radionuclide lung scanning with contrast echocardiography or AaPO^sub 2^ gradient in the supine position was also highly sensitive, with fewer than 5% false negative cases, and a predictive value of a negative result of 96%.

Estimates of Posttest Probability of Absence of PAVM in Patients with HHT

To determine the clinical utility of noninvasive tests as a screening procedure, we next calculated the posttest probability of the absence of PAVM in patients with HHT and negative diagnostic tests (Figure 2), using Bayesian equations based on the likelihood ratio of a negative test result and the pretest probability of the disease (21). The pretest probability of PAVM is represented by the prevalence of PAVM in the HHT population, which has been estimated as between 20 and 35% in previous studies depending on the screening method and the study population (7), and 48% in the present series. As shown in Figure 2, negative contrast cchocardiography excluded the diagnosis of PAVM with a probability of 97% in a population with a low prevalence of PAVM (20%), and a lower probability of 88% for an estimated prevalence of 50%. The combination of anterior chest radiograph and contrast echocardiography in the supine position excluded the diagnosis of PAVM with a probability of 100%.

DISCUSSION

Given the high risk of cerebral abscesses and ischemic complications of PAVM that may be prevented by embolization therapy, screening of patients with HHT for asymptomatic PAVM is warranted. Although several screening algorithms have been used (7, 12), none has been validated because tests commonly used for screening have not been compared with a gold standard (13). In the present study, we compared the diagnostic value of noninvasive tests for treatable PAVM (i.e., PAVM amenable to embolization therapy) in patients with HHT, using chest CT and/or pulmonary angiography as a reference. Contrast echocardiography was the most sensitive test. In addition, 100% sensitivity and negative predictive value could be obtained when combining anteroposterior chest radiograph and contrast echocardiography.

The present study differs from previous ones in that a high proportion of the patients underwent several if not all diagnostic tests. All tests were performed in more than 75% of the patients, and reference tests were conducted in all patients, thus considerably limiting the risk of bias. Nevertheless, substantial changes in technology occurred over the study period. As in any retrospective study, this study may suffer some bias. In this group of patients from a pulmonary department specialized in orphan lung disorders, 26.7% had a history of neurologic manifestations, representing 49% of patients with PAVM, a high proportion that may reflect selection bias as patients with neurologic symptoms are probably more frequently referred for pulmonary screening than asymptomatic persons. A high proportion of the patients with PAVM did not complain of dyspnea, confirming that screening cannot reliably be based on this symptom (15). Thus, our findings strongly support a systematic and presymptomatic pulmonary screening of patients with HHT. Because genetic heterogeneity of the disease has been suggested and the prevalence of PAVM may be higher in patients with mutations of the HHT1 gene encoding endoglin (22, 23), pulmonary screening should be particularly recommended in patients with HHT who have a family history of PAVM and/or such a mutation diagnosed. However, PAVMs also occur in families with mutations of the HHT2 gene (3).

Transthoracic contrast echocardiography utilizes microscopic gas bubbles to visualize right-to-left shunting. Intracardiac shunting (with presence of contrast in the left atrium within one cardiac cycle of its appearance in the right atrium) can be differentiated from intrapulmonary shunting (with a delay of three to eight cardiac cycles). Contrast echocardiography was the most sensitive test in our study and thus may be the best suited for screening. A good sensitivity of contrast echocardiography had been suggested by previous studies, but no definite conclusion could be drawn in the absence of a gold standard test in the group of patients with negative contrast echocardiography (14, 15, 24). We found falsely negative contrast echocardiography in a single patient diagnosed with treatable PAVM on CT (2 of 106 in another study [14]), possibly reflecting operatordependent results, as well as variability of the shunt with time or position. Transesophageal contrast echocardiography is more sensitive than transthoracic echocardiography for the detection of intrapulmonary shunt (20), but it is more invasive and thus unwarranted in asymptomatic patients with HHT. With a sensitivity and a negative predictive value of 100%, the combination of contrast echocardiography and chest radiograph may represent a good screening procedure, with the advantage of large availability, noninvasiveness, lower cost, and much lower radiation exposure than chest CT. Indeed, the probability of absence of PAVM in a patient with HHT and negative contrast echography was estimated according to Bayesian equations as between 88 and 97%, depending on the prevalence of PAVM in the HHT population (20-50%), and would be close to 100% with combined negative results of chest radiograph and contrast echocardiography. Because the posttest probability of the disease is dependent on the pretest probability, a more precise knowledge of the prevalence of PAVM in HHT would be useful.

The sensitivity of chest radiograph (70%) was lower than that of contrast echocardiography, as previously noted (4). However, its specificity was excellent in patients with HHT, and PAVMs large enough to be seen on chest radiography usually warranted embolization. Similarly, the sensitivity of radionuclide lung scanning was only 71%, consistent with a previous study in which this test was used to detect residual shunting after treatment of PAVM (25), so that this radioisotope diagnostic test is not indicated in a screening program. Measurement of right-to-left shunting by the 100% oxygen method is quantitative, and sensitivity and specificity are dependent on the cutoff values chosen. Thresholds of AaPO^sub 2^ based on optimal accuracy from a receiver-operator-characteristic curve were not well suited for screening, because sensitivity was only 80%. Using normal values of AaPO^sub 2^ from our laboratory, corresponding to a right-to-left shunt exceeding 5% of cardiac output (19), obtained excellent specificity (98%) but low sensitivity. Low thresholds of AaPO^sub 2^ had to be chosen to obtain a sensitivity of AaPO^sub 2^ of 90% or higher, and specificity dropped to 65%. We found this method less sensitive than was reported in previous studies, in which the sensitivity of the test was likely to be overestimated because only patients with positive contrast echocardiography were evaluated (4, 15). Given the fact that AaPO^sub 2^ measurement requires arterial puncture, and that its reproducibility among centers may be impaired by potential pitfalls (7), this method is not well suited for screening. Thus, we consider that radionuclide lung scanning and measurement of AaPO^sub 2^ should be abandoned as screening tests for PAVM.

Contrast echocardiography was the only positive test in 24 patients, none of whom was hypoxemic. For ethical reasons, patients with a negative CT scan were not subjected to an invasive procedure such as pulmonary angiography, the sensitivity of which, in any case, was lower than that of CT scan for treatable PAVM in the only available study comparing the two tests (16). cases with positive contrast echocardiography in patients with no visible PAVM correspond to false positive results only with respect to the need for embolization therapy, but microscopic and diffuse vascular dilatation causing right-to-left shunting might be present in such cases (15, 26). This hypothesis is consistent with the finding that transthoracic contrast echocardiography remains positive after endovascular treatment of PAVM, even in patients with no residual PAVM seen on angiography (27). The clinical significance of such possible microscopic PAVM and/or vascular dilatations is unknown. However, because it is recommended that patients with functional PAVM receive antibiotics before potentially bacteremic procedures to reduce the risk of brain abscess (7), we also propose such prophylactic treatment in patients with no visible PAVM but positive contrast echocardiography, although this has not been validated. Such patients may also be offered a closer clinical follow-up with chest CT after several years, to detect hemodynamically significant PAVM that may have evolved over time. In addition, further studies are required to determine whether grading of positive contrast echocardiography results might help to predict the development of significant PAVM (27).

In conclusion, this study provides evidence that transthoracic contrast echocardiography is the most sensitive noninvasive test for the screening of PAVM in patients with HHT. Screening with assessment of self-reported dyspnea, chest radiograph alone, measurement of AaPO^sub 2^ by the 100% oxygen method, or radionuclide lung scanning, was shown to be insufficient and may be abandoned. We suggest a screening algorithm (Figure 3) based on the combined use of contrast echocardiography and anteroposterior chest radiograph, followed by chest CT if either test is positive, and eventually pulmonary angiography in the case of PAVM amenable to embolization therapy with embolization done during the same procedure. Screening based only on chest CT might be an alternative to this algorithm. However, this algorithm may obviate the need of chest CT in patients without PAVM (i.e., 65-80% of patients with HHT). The proposed diagnostic algorithm needs to be validated in a prospective study, which will be conducted in our institution.

References

1. Shovlin CL, Guttmacher AE, Buscarini E, Faughnan ME, Hyland RH, Westermann CJ, Kjeldsen AD, Plauchu H. Diagnostic criteria for hereditary hemorrhagic telangiectasia (Rendu-Osler-Weber syndrome). Am J Med Genet 2000;91:66-67.

2. Plauchu H, de Chadarevian JP, Bideau A, Robert JM. Age-related clinical profile of hereditary hemorrhagic telangiectasia in an epidemiologically recruited population. Am J Med Genet 1989;32:291-297.

3. Shovlin CL, Letarte M. Hereditary haemorrhagic telangiectasia and pulmonary arteriovenous malformations: issues in clinical management and review of pathogenic mechanisms. Thorax 1999;54:714-729.

4. Haitjema T, Disch F, Overloom TT, Westermann CJ, Lammers JW. Screening family members of patients with hereditary hemorrhagic telangiectasia. Am J Med 1995;99:519-524.

5. Swinburne AJ, Fedullo AJ, Gangemi R, Mijangos JA. Hereditary telangiectasia and multiple pulmonary arteriovenous fistulas: clinical deterioration during pregnancy. Chest 1986;89:459-460.

6. Maher CO, Piepgras DG, Brown RD Jr, Friedman JA, Pollock BE. Cerebrovascular manifestations in 321 cases of hereditary hemorrhagic telangiectasia. Stroke 2001;32:877-882.

7. Gossage JR, Kanj G. Pulmonary arteriovenous malformations: a stale of the art review. Am J Respir Crit Care Med 1998;158:643-661.

8. Moussouttas M, Fayad P, Rosenblatt M, Hashimoto M, Pollak J, Hender son K, Ma TY, White RI. Pulmonary arteriovenous malformations: cerebral ischemia and neurologic manifestations. Neurology 2000;55:959-964.

9. Press OW, Ramsey PG. Central nervous system infections associated with hereditary hemorrhagic telangiectasia. Am J Med 1984;77:86-92.

10. While RI Jr, Lynch-Nyhan A, Terry P, Buescher PC, Farmlett EJ, Charnas L, Shuman K, Kim W, Kinnison M, Mitchell SH. Pulmonary arteriovenous malformations: techniques and long-term outcome of embolotherapy. Radiology 1988;169:663-669.

11. Gupta P, Mordin C, Curtis J, Hughes JM, Shovlin CL, Jackson JE. Pulmonary arteriovenous malformations: effect of embolization on right-to-left shunt, hypoxemia, and exercise tolerance in 66 patients. AJR Am J Roentgenol 2002;179:347-355.

12. Hughes JM. Pulmonary arteriovenous malformations in hereditary hemorrhagic telangiectasia. Semin Respir Crit Care Med 1998; 19:79-89.

13. Gossage JR. The role of echocardiography in screening for pulmonary arteriovenous malformations. Chest 2003;123:320-322.

14. Nanthakumar K, Graham AT, Robinson TI, Grande P, Pugash RA, Clarke JA, Hutchison SJ, Mandzia JL, Hyland RH, Faughnan ME. Contrast echocardiography for detection of pulmonary arteriovenous malformations. Am Heart J 2001;141:243-246.

15. Kjeldsen AD, Oxhoj H, Andersen PE, Elle B, Jacobsen JP, Vase P. Pulmonary arteriovenous malformations: screening procedures and pulmonary angiography in patients with hereditary hemorrhagic telangiectasia. Chest 1999;116:432-439.

16. Remy J, Remy-Jardin M, Wattinne L, Deffontaines C. Pulmonary arteriovenous malformations: evaluation with CT of the chest before and after treatment. Radiology 1992;182:809-816.

17. Cottin V, Frognier R, Deygas N, Gentil B, Plauchu H, Cordier JF. Pulmonary arteriovenous malformations: diagnostic value of screening tests in 40 patients with hereditary hemorrhagic telangiectasia [abstract]. Am J Respir Crit Care Med 2003;167:A695.

18. Quanjer PH, Tammeling GJ, Cotes JE, Pedersen OF, Peslin R, Yernault JC. Lung volumes and forced ventilatory flows. Eur Respir J 1993;6(Suppl 16):5-40.

19. Genero M, Abid A, Rezig M, Philit F, Bertocchi M, Mornex JF, Cordier JF, Brune J, Wiesendanger T. Alveolar-oxygen difference under pure oxygen for evaluation of right-to-left shunt. Chest 1996;110:1968.

20. Vedrinne JM, Duperret S, Bizollon T, Magnin C, Motin J, Trepo C, Ducerf C. Comparison of transesophageal and transthoracic contrast echocardiography for detection of an intrapulmonary shunt in liver disease. Chest 1997;111:1236-1240.

21. Weissler AM. A perspective on standardizing the predictive power of noninvasive cardiovascular tests by likelihood ratio computation: 1. Mathematical principles. Mayo Clin Proc 1999;74:1061-1071.

22. Porteous ME, Curtis A, Williams O, Marchuk D, Bhattacharya SS, Burn J. Genetic heterogeneity in hereditary haemorrhagic telangieclasia. J Med Genet 1994;31:925-926.

23. Berg JN, Guttmacher AE, Marchuk DA, Porteous ME. Clinical heterogeneity in hereditary haemorrhagic telangieclasia: arc pulmonary arteriovenous malformations more common in families linked to endoglin? J Med Genet 1996;33:256-257.

24. Oxhoj H, Kjeldsen AD, Nielsen G. Screening for pulmonary arteriovenous malformations: contrast echocardiography versus pulse oximetry. Scand Cardiovasc J 2000;34:281-285.

25. Thompson RD, Jackson J, Peters AM, Dore CJ, Hughes JMB. Sensitivity and specificity of radioisotope right-left shunt measurements and pulse oximetry for the early detection of pulmonary arteriovenous malformations. Chest 1999;115:109-113.

26. Kjeldsen AD, Oxhoj H, Andersen PE, Green A, Vase P. Prevalence of pulmonary arteriovenous malformations (PAVMs) and occurrence of neurological symptoms in patients with hereditary haemorrhagic telangiectasia (HHT). J Intern Med 2000;248:255-262.

27. Lee WL, Graham AF, Pugash RA, Hutchison SJ, Grande P, Hyland RH, Faughnan ME. Contrast echocardiography remains positive after treatment of pulmonary arteriovenous malformations. Chest 2003;123:351-358.

Vincent Cottin, Henri Plauchu, Jean-Yves Bayle, Martine Barthelet, Didier Revel, and Jean-Francois Cordier

Service de Pneumologie, Centre des Maladies Orphelines Pulmonaires, Hopital Cardiovasculaire et Pneumologique Louis Pradel, Universite Claude Bernard; Service de Genetique, Hopital de l'Hotel-Dieu; Laboratoire d'Exploration Fonctionnelle Respiratoire, Laboratoire d'Echocardiographie, and Service de Radiologie, Hopital Louis Pradel; Reseau de Recherche sur la Maladie de Rendu-Osler; and UMR 754 INRA-ENVL-UCBL, IFR 128, Lyon, France

(Received in original form October 23, 2003; accepted in final form January 75, 2004)

Supported by grants HCL-PHRC 93.97-005 and HCL-JCH 2002, and by INSERM-AFM-Ministere Francais de la Recherche 2000 (Reseau Maladies Rares 2000).

Correspondence and requests for reprints should be addressed to Vincent Cottin, M.D., Ph.D., Hopital Cardiovasculaire et Pneumologique Louis Pradel, 69677 Bron Cedex, France. E-mail: vincent.cottin@chu-lyon.fr

Am J Respir Crit Care Med Vol 169. pp 994-1000, 2004

Originally Published in Press as DOI: 10.1064/rccm.200310-1441OC on January 23, 2004

Internet address: www.atsjournals.org

Conflict of Interest Statement: V.C. has no declared conflict of interest; H.P. has no declared conflict of interest; J-Y.B. has no declared conflict of interest; MB. has no declared conflict of interest; D.R. has no declared conflict of interest; J-F.C. has no declared conflict of interest.

Acknowledgment: The authors thank Dr. C. Houzard (Lyon), who performed the radionuclide lung scanning. The authors thank Prof. F. Gueyffier and Prof. R. Ecochard for methodologic contribution. The authors thank Drs. N. Deygas, B. Gentil, and B. Etienne-Mastroianni for clinical pulmonary evaluation of the patients, and Drs. E. Servan, M. O. Rial, and F. Jamal for performing echocardiography. The authors thank Prof. J. Honnorat, Prof. C. Pignat, and Dr. J. C. Saurin for clinical evaluation of the patients in the regional network on HHT. The authors thank Drs. R. Frognier, A. S. Blanchet, and J. O. Maillard for data entry.

Copyright American Thoracic Society May 1, 2004
Provided by ProQuest Information and Learning Company. All rights Reserved