| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Imaging, Imperial College Faculty of Medicine, Hammersmith
Hospital, Du Cane Rd., London W12 0NN, England.
2 Department of Respiratory Medicine, National Heart and Lung Institute,
Imperial College Faculty of Medicine, Hammersmith Hospital, London W12 0NN,
England.
Received October 1, 2001;
accepted after revision February 14, 2002.
Address correspondence to J. E. Jackson.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. In 58 (88%) of 66 patients, all malformations with feeding vessels greater than or equal to 3 mm in diameter were embolized with steel coils. Arterial oxygen saturation at rest and exercise, intrapulmonary right-to-left anatomic shunt fraction (99mTc-macroaggregate injection), maximum exercise capacity (incremental work rate test), and pulmonary function were measured before and after embolization. Complications were analyzed.
RESULTS. Three categories of patients were identified. Patients in group 1 (27%) had complete occlusion of all angiographically visible pulmonary arteriovenous malformations; patients in group 2 (61%) had complete occlusion of all malformations with feeding vessels greater than or equal to 3 mm in diameter, but with smaller lesions persisting; and patients in group 3 (12%) had incomplete embolization, with feeding vessels greater than or equal to 3 mm in diameter remaining. The mean right-to-left shunt after embolization was least in group 1 (7%), intermediate in group 2 (10%), and greatest in group 3 (19%). Arterial oxygen saturation and right-to-left shunt fraction returned to normal levels (>96% and <3.5%, respectively) in 33% of patients. A significant improvement occurred after embolization in carbon monoxide diffusing capacity per unit of alveolar volume and in exercise capacity in 16 and 10 patients, respectively. In 93 procedures, 12 complications (13%) occurred.
CONCLUSION. Coil embolization of pulmonary arteriovenous malformations is effective in reducing right-to-left anatomic shunt fraction and in improving arterial oxygenation. Coil embolization of pulmonary arteriovenous malformations is well tolerated and has a low complication rate.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Complications relating to pulmonary arteriovenous malformations are common [2]; most can be attributed to the intrapulmonary right-to-left shunt, which in severe cases may exceed 40% of the cardiac output. Hypoxemia is common but is usually well tolerated by the patient because of the attendant low pulmonary vascular resistance. Nevertheless, cerebral abscesses, cerebrovascular accidents, and transient ischemic attacks due to paradoxic embolism through the pulmonary arteriovenous malformations result in considerable morbidity and mortality [8,9,10,11]. Although individual pulmonary arteriovenous malformations were once treated by surgical resection, multiple lesions are likely to develop over time in most individuals with hereditary hemorrhagic telangiectasia. Transcatheter embolization has transformed the outlook for patients and is now considered the treatment of choice [12]. To reduce the incidence of paradoxic embolism, all pulmonary arteriovenous malformations with feeding vessels amenable to endovascular occlusion should be so treated, and antibiotic prophylaxis should be administered for dental and surgical procedures in patients with residual shunts [2, 13].
Articles about embolization provide evidence for regression of the pulmonary arteriovenous malformation sac [14], reduction in the right-to-left shunt [15,16,17,18,19,20], and improvement in hypoxemia [2, 15, 18,19,20,21,22,23,24]. However, published series also reveal varying efficacy and complications of treatment, according to the center where the study was performed. In this article, we present 66 patients who have not been described in any study. This series presents new therapeutic protocols and highlights further improvement in the efficacy of treatment and reductions in the incidence of embolization-associated complications. We have again shown a significant improvement in KCO (DL/VA) (the diffusing capacity for carbon monoxide per unit of alveolar volume) in a subgroup of treated patients in whom KCO before embolization was less than 90% of the predicted value. In addition, we have introduced a new categorization for pulmonary arteriovenous malformation patients to provide a suitable framework for longterm follow-up studies.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Pulmonary Function Tests
Standard pulmonary function tests were performed, including forced
expiratory volume in 1 sec (FEV1), vital capacity, single-breath
diffusing capacity for carbon monoxide (DLCO), and diffusing
capacity for carbon monoxide per unit of alveolar volume (KCO
[DL/VA]). The mean values for these measurements were corrected for
the current hemoglobin concentration and expressed as a percentage of the
predicted value.
Arterial oxygen saturation (SaO2) was measured with the patient breathing room air using a pulse oximeter (Biox 3740; Ohmeda, Hatfield, Hertfordshire, UK) and an ear probe in the standing and the supine positions (each for 10 min, with recordings every 60 sec and the result expressed as the mean of the last four readings). SaO2 was also measured at maximal exercise on a cycle ergometer (Ergomed 840 cycle ergometer; Siemens, Bracknell, Berkshire, United Kingdom) using a 15-30 W/min incremental protocol. For patients unable to cycle, the distance walked on a level surface in 3 min was measured with a recording of SaO2 and pulse rate at the end of the walk.
Right-to-Left Shunt Assessment
Right-to-left shunt was measured by the IV injection of
99mTc-labelled albumin macroaggregates followed by gamma camera
imaging. The principle and method of this technique have been described
previously [25,
26]. In brief, calculations of
the shunt fraction were made using the counts in the right kidney, corrected
for attenuation, and expressed as a percentage of the injected dose, with the
right-to-left shunt expressed as 10 times this value assuming the right kidney
received 10% of the cardiac output (kidney-dose method). A second method
calculated the right kidney counts as a proportion of the counts over the
lungs (kidney—lung method). For each subject, the mean of the
measurements obtained using kidney-dose and kidney—lung methods was
used. For some patients, multiple measurements were recorded on different
occasions. The mean right-to-left shunt measurement before embolization was
taken from the average of the earliest recorded results; the mean shunt
measurement after embolization was taken from the average of the latest
recorded results.
Embolization Procedure
The same person performed all pulmonary embolizations. Prophylactic
antibiotics (500 mg of vancomycin) were administered 1 hr before and 8 hr
after each procedure. Diagnostic pulmonary arteriograms preceded pulmonary
arteriovenous malformation embolization in a single session. If required,
patients were readmitted for further elective embolization after approximately
3 months; each embolization procedure lasted approximately 2-2.5 hr.
Bilateral pulmonary arteriography in frontal and oblique projections was initially performed in all patients after measurement of the pulmonary artery pressure. The pigtail catheter used for these arteriograms was then exchanged for a Head-hunter I catheter (Cordis; Ascot, Berkshire, United Kingdom), and embolization of the feeding vessels to the individual pulmonary arteriovenous malformations was achieved with metallic coils.
The technique of embolization has been described [27] and involves placing coils as close as possible to the neck of the malformation in an attempt to avoid the occlusion of normal pulmonary artery branches. Coils of a size depending on the diameter of the feeding artery were deployed in each vessel until complete occlusion was obtained. Most of the coils used in this series were mechanically detachable (Jackson Detachable Coil; William Cook Europe, Bjaeverskov, Denmark) [28]. In three patients, packing of the venous sac with coils [29] was performed because the anatomy of the arteriovenous malformations in these individuals precluded conventional embolization; in another two patients, embolization of two large high-flow pulmonary arteriovenous malformations was achieved with coils inserted through the lumen of a standard occlusion balloon catheter that was temporarily inflated to obliterate flow in the feeding vessel.
Follow-Up
Patients were followed up in a dedicated pulmonary arteriovenous
malformation clinic 3 months after their final embolization and annually
thereafter. Clinical examination, pulmonary function tests, and radio-nuclide
studies were performed. The average period of follow-up after embolization was
27.3 months (range, 0-67 months).
Statistical Analysis
The data were analyzed using the statistical software STATA (version 6;
Stata, College Station, TX). To identify associations in the data, we used the
chisquare test for categoric variables. The nonparametric Wilcoxon's matched
pairs signed rank test was used to compare continuous measurements. To
compensate for multiple comparison tests, only p values of less than
0.01 were considered statistically significant.
Values obtained before and after embolization do not include repeated measurements taken of individuals who underwent multiple procedures. The value before embolization corresponds to the subject's earliest recorded value and the value after embolization corresponds to the subject's latest recorded value, resulting in only one set of values per subject.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Although hypoxemia provides a strong erythropoietic stimulus, in many patients the resultant secondary polycythemia was masked by concurrent nasal or gastrointestinal hemorrhage caused by underlying hereditary hemorrhagic telangiectasia. Eighteen patients (27%) were polycythemic (male, hemoglobin >17.7 g/dL; female, hemoglobin >15.2 g/dL) at some stage. Three patients required venesection. Polycythemia was seen in 33% (15/45) of females and 14% (3/21) of males.
Ten patients received blood transfusions. No significant sex difference was noted in the incidence of anemia (females: 13/45, 29%; males: 9/21, 43%; chi-square test = 1.26, p = 0.262).
Embolization
Ninety-three procedures were performed in 66 patients (range, one to four
procedures; mean, 1.4 procedures per patient). Forty-eight patients underwent
a single procedure, and 18 patients underwent multiple procedures. A total of
225 pulmonary arteriovenous malformations (right lung, 125 [56%]; left lung,
100 [44%]) were embolized. The zonal distribution of the embolized lesions was
11% in the upper zone, 29% in the mid zone, and 60% in the lower zone. A mean
of 2.5 pulmonary arteriovenous malformations (range, one to seven
malformations) were occluded with an average of 10 metallic coils (range, one
to 39 coils) per procedure. The feeding vessels ranged from 2 to 12 mm in
diameter (mean, 6 mm). Mean pulmonary artery pressure before embolization was
15.3 mm Hg (range, 8-36 mm Hg), and mean systolic and diastolic pressures were
26 and 8.8 mm Hg, respectively.
The patients were divided into three groups. Group 1 patients (n = 18, 27%) experienced complete occlusion of all angiographically visible pulmonary arteriovenous malformations. Group 2 patients (n = 40, 61%) experienced complete occlusion of all pulmonary arteriovenous malformations with feeding vessels greater than or equal to 3 mm diameter, but with smaller lesions persisting on angiography. In group 3 patients (n = 8, 12%), embolization was incomplete and feeding vessels larger than 3 mm remained. Four patients are awaiting further embolization, and one has since died of unrelated causes. In the remaining three group 3 patients, embolization was incomplete because of increasing pulmonary artery pressures (n = 1) or because of the risk of occluding pulmonary artery branches that supply large amounts of normal lung (n = 2).
One potentially serious complication occurred in the 93 procedures when a 5-mm coil migrated through the pulmonary arteriovenous malformation into the left popliteal artery, from which it was successfully snared and removed without any sequelae. Five patients experienced mild central chest pain during the procedure; in four patients this pain was accompanied by elevation of the ST segment on an EKG, and the pain was rapidly relieved by sublingual nitrate and 100% oxygen. Another patient had mild chest discomfort, most likely musculoskeletal in origin, that was relieved by 100% oxygen. In two patients, a small part of the normal lung parenchyma had to be sacrificed. After embolization, three patients experienced pleuritic chest pain that resolved with nonsteroidal antiinflammatory agents. One patient had a small hemoptysis associated with an area of alveolar shadowing that occurred at the lung base during embolization and that was probably caused by the rupture of a small vessel during coil occlusion. The patient was well and pain-free the next day and has experienced no further hemoptysis.
After Embolization
After maximal embolization, significant improvement was seen in the
SaO2 in both postures, on maximal exercise, and in the
right-to-left shunt in each group (Tables
2,3,4,5).
Overall, using our previously published criteria
[37], the right-to-left shunt
returned to normal (<3.5%) in 22 patients (33%); erect SaO2 was
greater than 96% in 28 patients (42%). A significant improvement occurred in
SaO2 on maximal exercise testing (p < 0.0001) overall,
but no significant improvement in workload was achieved in any of the three
groups. However, for groups 1 and 2 combined, a significant difference
(p = 0.004) was noted between the workload measurements before and
after embolization, although only 10 (22%) of 46 patients actually improved,
and the maximum improvement observed was only 45 W.
|
|
|
|
No significant change was seen in the mean FEV1, vital capacity, DLCO, and KCO after embolization (Table 6). In the subgroup of patients (n = 16) with KCO before embolization of less than 90% of the predicted value, all except two showed improvement after embolization. In these 16 patients, embolization resulted in a statistically significant improvement in KCO (p = 0.004, Wilcoxon's matched pairs signed rank test).
|
Follow-Up
Pulmonary arteriovenous malformations.—Three patients had
successful pregnancies after embolization of pulmonary arteriovenous
malformations. All these patients experienced considerable deterioration in
right-to-left shunt, arterial oxygenation, and exercise capacity postpartum.
Angiograms obtained after delivery showed enlargement of the residual lesions
that were successfully embolized in two patients. In one patient, the
pulmonary arteriovenous malformations were considered too small (<3 mm in
diameter) and too numerous for further embolization.
General.—Two deaths occurred, both unrelated to pulmonary arteriovenous malformations. One patient had a cardiorespiratory arrest after appendicular surgery. In the second patient, a postmortem examination revealed pneumonia. One patient with a cerebral arteriovenous malformation developed an intracerebral hematoma during follow-up that led to considerable disability.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Good though present treatment is, small residual lesions remain in 60% of patients (Table 4, group 2). These lesions stem from pulmonary arteriovenous malformations whose feeding vessels are too small (<3 mm in diameter) to be embolized with current technology. These patients continue to be at risk from paradoxic embolization and neurologic sequelae, although the risk is significantly less than it was before treatment. Nevertheless, antibiotic prophylaxis before dental procedures and surgical treatment should continue.
Interventional Techniques
By far most pulmonary arteriovenous malformations can be successfully
treated by conventional embolization whereby coils are placed at the neck of
the malformation immediately proximal to the dilated aneurysmal sac (Fig.
1A,1B,1C).
It is important to achieve a distal occlusion, if possible, for the following
reasons: First, vessels that arise from the feeding vessel that supplies the
normal lung, which may not be visualized on arteriography performed before
embolization because of a steal effect through a high-flow shunt, will be
preserved (Fig.
1A,1B,1C),
and this portion of the lung will maintain its ability to participate in gas
exchange. Although preserving vessels may not be clinically important in
patients with one or two pulmonary arteriovenous malformations whose remaining
lung is otherwise normal, it is likely to be relevant in those who have
numerous shunts and a reduced amount of normal pulmonary parenchyma.
Preservation of these branches may be one of the factors that reduces the
incidence of pleuritic chest pain after embolization, which has been reported
to occur in 10% or more of individuals undergoing pulmonary arteriovenous
malformation embolization in some series
[18,
21]. Our incidence of this
complication was only 3%, which may reflect the great importance we attach to
distal embolization. Second, distal occlusion may reduce the likelihood of
continued perfusion of the venous sac by normal bronchopulmonary anastomoses,
although the clinical importance of this is unknown.
|
|
|
Packing the venous sac.—Only five of 225 pulmonary arteriovenous malformations could not be treated by conventional coil occlusion; three were embolized by packing the venous sac with coils (Fig. 2A,2B,2C). This technique was required when a very short or wide neck precluded embolization because of the risk of occlusion of large pulmonary artery branches that supply normal lung. Coils of appropriate size were therefore placed in the venous sac to provide a frame on which smaller coils could be used to completely occlude the neck. This form of embolization is not appropriate for most pulmonary arteriovenous malformations because of the theoretic risk of sac rupture and an increased possibility of migration of thrombi that form on these coils into the draining pulmonary vein and thence into the systemic circulation. In addition, the resulting "coil ball" will remain indefinitely, whereas in a pulmonary arteriovenous malformation treated by conventional coil occlusion, the sac will usually disappear (Fig. 3A,3B,3C,3D,3E,3F).
|
|
|
|
|
|
|
|
|
Use of an occlusion balloon.—In two patients with especially large pulmonary arterio-venous malformations through which the flow was extremely rapid, an occlusion balloon was inflated in the feeding vessel to control flow during coil deployment, which was performed through the lumen of the occlusion balloon catheter (Fig. 3A,3B,3C,3D,3E,3F). This technique is useful when a stable catheter position at a site appropriate for embolization in large feeding vessels cannot be achieved with a conventional catheter. The two lesions treated in this manner had feeding vessels of 7 and 9 mm in diameter.
Coil Migration
A single potentially serious complication occurred when a coil that was too
small for the vessel being embolized migrated through a pulmonary
arteriovenous malformation into the systemic circulation and lodged in the
left popliteal artery, from which it was successfully retrieved. Coil
migration has not occurred again from the time that we began using detachable
coils routinely. Coil migration remains a possible complication that is
discussed with all patients before the procedure. In the worst scenario, the
coil migrates into the cerebral vessels and causes cerebrovascular ischemia
and perhaps infarction. We know of only one instance of coil migration to an
internal carotid artery, from which it was successfully retrieved without
complication (White R, personal communication).
Gas Exchange
Successive series from our institution
[16,
18] and others
[15,
17,
19] have shown a decrease in
the right-to-left shunt both before and after embolization. A similar
improvement was noted in all indexes of oxygenation (SaO2) and, in
a subgroup of patients with impaired gas transfer, improvement in
KCO. The probable reason for the higher SaO2 and lower
right-to-left shunt before embolization in more recent series is that family
screening procedures and better awareness of hereditary hemorrhagic
telangiectasia and its complications by patients, their families, and
physicians, mean that patients are now presenting at a less symptomatic stage
and before serious complications occur. Tables
3,4,5
summarize the physiologic data in terms of oxygenation, right-to-left shunt,
and exercise capacity for three categories of patients: group 1, patients with
complete occlusion of all angiographically visible pulmonary arteriovenous
malformations; group 2, patients with residual shunts remaining in pulmonary
arteriovenous malformations with feeding vessels too small for embolization
therapy (<3 mm in diameter); and group 3, patients whose treatment was
incomplete and in whom pulmonary arteriovenous malformations with feeders
greater than 3 mm remain. The mean shunt after embolization was least in group
1 (7%), intermediate in group 2 (10%), and greatest in group 3 (19%). Tables
3,4,5
also show a similar trend among groups in all indexes of oxygenation.
Interestingly, in group 1, whose treatment was angiographically complete, the
anatomic right-to-left shunt remained abnormal by our standards
[37] at greater than 3.5% in
two (11%) of 18 patients, implying (in the absence of a patent foramen ovale)
that pulmonary arteriovenous malformations beyond the resolution of
angiography must have been present.
Exercise Outcomes
Most patients subjectively report an improvement after embolization in
exercise capacity in terms of daily living
[15], a finding with which we
agree; an objective increase in exercise capacity (increase in the oxygen
uptake at maximal exercise [Vo2(max)]) has also been reported in a
small number of patients [17].
In this series, and in two previous reports from our hospital, we found that
an objective increase in exercise capacity occurs in only a minority of
patients after embolization (6/14
[23], 15/41
[18], and 10/46 in this
study). This finding has two probable reasons. First, the exercise protocol,
generally an increase in workload of 30 W/min, is a somewhat blunt tool that
may overlook subtle but small improvements. Second, patients with pulmonary
arteriovenous malformations have efficiently adapted to delivering oxygen to
muscles that are exercising (polycythemia and a supranormal cardiac output)
[38], so that their exercise
capacity before embolization may not be markedly impaired.
Pulmonary Function
In the absence of pulmonary infarction and scarring, no reason exists for
the FEV1 and vital capacity to change after embolization; this has
been the case in this and other series. Because DLCO and
KCO reflect the function of the pulmonary microcirculation, they
might not be expected to be affected by embolization of large pulmonary
arteriovenous malformations. In two other series
[17,
23], the researchers did not
find any significant increase in DLCO or KCO except in a
subgroup of four patients with pulmonary arteriovenous malformations that were
not associated with hereditary hemorrhagic telangiectasia
[17]. On the other hand, we
showed in another series of patients that a small (1.3%) insignificant
increase in DLCO and a small but significant increase in
KCO (5.2%; p = 0.02) occurred after embolization
[18]. Most patients in our
study had DLCO and KCO within the normal range before
embolization, and we specifically analyzed the subgroup (16/60) whose
KCO before embolization was less than 90% of the predicted value.
In that group, we found a significant improvement in KCO, from
77.5% to 86.8%.
The subgroup with improved KCO does not correspond to the subgroup with improved exercise capacity. Of the patients in whom exercise capacity improved, only three of 13 had KCO of less than 90% of the predicted value. Conversely, exercise tolerance increased in only three of 13 patients with KCO of less than 90% for whom full exercise measurements were available, and only one of these had an increase in KCO as a result of the procedure. Therefore, the improvements in exercise tolerance and in DLCO in subgroups of patients appear to be independent. In general, impaired gas transfer is associated with impaired exercise capacity, so this finding is a little surprising and indicates the extent to which additional factors such as hemodynamics may influence exercise capacity.
Conclusion
Coil embolization of pulmonary arteriovenous malformations is a
well-tolerated procedure with a low complication rate and is effective in
terms of reduction in right-to-left anatomic shunt fraction and improvement in
arterial oxygenation. Although most patients report a subjective improvement
in exercise tolerance after treatment, an objective improvement is difficult
to show. In individuals with a KCO of less than 90% of the
predicted value, a significant improvement of this measurement can be expected
after embolization.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Gazzaniga, E. Buscarini, G. Leandro, L. Reduzzi, M. Grosso, G. Pongiglione, C. Pedrinazzi, L. Lanzarini, V. Portugalli, P. Blotta, et al. Contrast echocardiography for pulmonary arteriovenous malformations screening: does any bubble matter? Eur J Echocardiogr, December 17, 2008; (2008) jen317v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. A. Ali, M. Lippmann, U. Mundathaje, and G. Khaleeq Spontaneous Hemothorax: A Comprehensive Review Chest, November 1, 2008; 134(5): 1056 - 1065. [Full Text] [PDF]
|
|
|
|
 |
|
 C. L. Shovlin, H. C. Tighe, R. J. Davies, J. S. R. Gibbs, and J. E. Jackson Embolisation of pulmonary arteriovenous malformations: no consistent effect on pulmonary artery pressure Eur. Respir. J., July 1, 2008; 32(1): 162 - 169. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Schneider, M. Uder, M. Koehler, M. A. Kirchin, A. Massmann, A. Buecker, and U. Geisthoff MR Angiography for Detection of Pulmonary Arteriovenous Malformations in Patients with Hereditary Hemorrhagic Telangiectasia Am. J. Roentgenol., April 1, 2008; 190(4): 892 - 901. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C L Shovlin, J E Jackson, K B Bamford, I H Jenkins, A R Benjamin, H Ramadan, and E Kulinskaya Primary determinants of ischaemic stroke/brain abscess risks are independent of severity of pulmonary arteriovenous malformations in hereditary haemorrhagic telangiectasia Thorax, March 1, 2008; 63(3): 259 - 266. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Poster presentations Thorax, December 1, 2006; 61(suppl_2): ii57 - ii133. [Full Text] [PDF]
|
|
|
|
 |
|
 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]
|
|
|
|
 |
|
 A J Easey, G M F Wallace, J M B Hughes, J E Jackson, W J Taylor, and C L Shovlin Should asymptomatic patients with hereditary haemorrhagic telangiectasia (HHT) be screened for cerebral vascular malformations? Data from 22 061 years of HHT patient life J. Neurol. Neurosurg. Psychiatry, June 1, 2003; 74(6): 743 - 748. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||
