Pulmonary Arteriovenous

Aneurysms and Fistulas

Anatomical Variations, ErnbryologJ7,
and Classification

Isam N. Anabtawi, M.D., Robert G. Ellison, M.D.,

and Lois T. Ellison, M.D.

T he clinical and pathological findings of pulmonary arterio-

venous fistulas were first correlated by Rodes [16]. The first
successful surgical treatment of this condition was reported by
Shenstone [17] in 1942 in a case which was diagnosed by Hepburn and
Dauphinee [7]. Since then numerous articles on the subject have ap-
peared in the literature. Comprehensive reviews have been published
by Sloan and Cooley [ 181 and Le Roux [9].
The etiology of these anomalies has been a subject of conjecture.
Their relationship to physiological arteriovenous shunts and hemangi-
omatous malformations seen elsewhere in the body suggests abnormal
development of the capillary system. Variations in anatomical structure,
arterial supply, and venous drainage have also been reported. How-
ever, no attempt has been made to correlate these anomalies with the
embryological development of the pulmonary vascular system. T h e
relationship of these fistulas to anomalous venous drainage has also
escaped at ten tion.
The purpose of this paper is to attempt a classification of these
anomalies and to point out their relationship with anomalous venous
drainage. A brief review of the embryology of the pulmonary vascular
system will shed light on the encountered variations.

EMBRYOLOGY
Early in embryological life the respiratory system arises as a median
ventral diverticulum of the foregut. A longitudinal groove is present
on either side which gradually deepens so that on further growth the
From the Department of Surgery (Thoracic) and the Hemotlynamic Research Unit, Medical
College of Georgia, and Eugene Talmadge Memorial Hospital, Augusta, Ga.
Supported by Public Health Service Research Grant Nos. HE-07266 and HE-2562.
Received for publication Nov. 2, 1964.

VOL. 1, NO. 3, MAY, 1965 277

ANABTAWI, ELLISON, AND ELLISON

respiratory primordium becomes more separated from the esophagus.

Each endodermal outgrowth which constitutes the primitive lung bud
is surrounded by a venous plexus which is an extension of the foregut
plexus of veins. T h i s plexus of veins anastomoses with arborization of
the primitive pulmonary arteries arising from the sixth aortic arch.
These arteriovenous connections can be regarded as multiple small
arteriovenous fistulas. These have been well demonstrated in the human
embryo and fetus [20, 211. With the passage of time further elaboration
occurs. Vascular septa develop and capillaries are formed, while con-
fluence of some veins results in the formation of larger vessels.
T h e pulmonary veins develop from the confluence of venous
channels, forming a single large venous trunk in close proximity to the
heart. This trunk dilates proximally and is absorbed into the left
atrium. Distally, the vein dilates further and two trunks develop.
T h e trunks move apart to lie on either side of the left atrium so that
a portion of this trunk contributes to the formation of a portion of the
posterior wall of the left atrium.
T h e large venous trunk is in close proximity to the sinus venosus,
so that failure of absorption into the left atrium causes this vein to
absorb into the sinus venosus, thus forming an anomalous drainage
into the right atrium (sinus venosus type of ostium secundum). It is
uncertain whether an abnormal position of the septum contributes to
this anomaly. Failure of absorption into either atrium results in the
persistence of the large venous trunk posterior to the left atrium. T h e
blood is then diverted via the cardinal system of veins into the superior
vena cava or inferior vena cava (total anomalous venous drainage).
From the foregoing description it is conceivable that abnormalities
of development might occur at the pulmonary arterial, capillary, or
venous level as isolated or combined lesions, which would justify the
following classification. Evidence in support of this classification will
be discussed.

M A L D E V E L O P M E N T S OF P U L M O N A R Y V A S C U L A R SYSTEM
A schematic presentation of the encountered variations is depicted
in Figure 1.

MULTIPLE SMALL ARTERIOVENOUS FISTULAS WITHOUT ANEURYSMS

T h e existence of pulmonary arteriovenous shunts in normal states
has been verified by several investigators. Gas studies show a certain
degree of venous admixture in the absence of intracardiac shunts 15,
12, 141. T h e passage of glass or plastic beads, twenty to forty times the
size of pulmonary capillaries, and certain parasites, such as Schistosoma
cercariae [ 191, through the pulmonary vascular bed is further proof

278 THE ANNALS OF THOKACIC SUKGEKY

 U I. II. Ill A.
Ill B.
Multiple Small A-V Fistulas Large A-V Aneurysm Large A-V Aneurysm Large A-V Aneurysm with
without Aneurysm (peripheral 1 k e ntra II Anomalous Venous Drainage
e@
W t
5 I V B.
Ill c. V.
Multiple Small A-V Fistulas Large Venous Aneurysm Large Venous Aneurysm Anomalous Venous Drainage
with Anomalous Venous with Systemic Artery without Fistula without Fistulas
Drainage Communication (Undescribed)

F I G . 1. Proposed anatomical classification of pulmonary arteriovenous fistulas and anwrysms.

ANABTAWI, ELLISON, AND ELLISON

of the existence of such shunts. Recent experimental studies indicate

that normally these pulmonary arteriovenous shunts measure 35y to 50p
in diameter in the dog [25]. Physiological studies indicate that these
shunts enlarge when pulmonary artery systolic pressure exceeds 40 mm.
H g El.
These shunts are mainly located in the apices of the segmental
subdivisions and in close proximity to the visceral pleura. This an-
atomical distribution is similar to that of pathological arteriovenous
fistulas. It is interesting that embryologically the last vascular differ-
entiation takes place in these locations. Hemodynamic studies have
also shown that abnormally large shunts develop in congenital heart
diseases accompanied by large left-to-right shunts or pulmonary hy-
pertension [8]. I n other pulmonary hypertensive states we have demon-
strated abnormal venous admixture which may be due to enlarged
arteriovenous shunts. T h e role of these shunts in protection of the
pulmonary vascular bed is suggested.
I n 1951 Brink [3] reported a case of multiple hereditary telan-
giectasis with cyanosis and clubbing. T h e pulmonary veins were not
demonstrable by radiographic means, including angiocardiography.

A B
FIG. 2. ( A ) Posteroanterior chest film of a 2-year-old white female with multiple
small arteriouenous fistulas not demonstrable by angiocardiography. ( B ) Hands
and feet of same patient showing marked clubbing and cyanosis.

280 THE ANNALS OF THORACIC SURGERY

 Pulmonary Arteriouenous Aneurysms and Fistulas

A similar case was reported by Cooley and McNamara [4] in 1954, and
the diagnosis was confirmed by exploratory thoracotomy and biopsy
for the first time. This case was thought to represent multiple hereditary
telangiectasis. Recently we observed a moderately cyanotic 2-year-old
white girl who had cyanosis and clubbing since infancy (Fig. 2). Ex-
tensive studies, including cardiac catheterization, angiocardiography,
and dye-dilution studies were suggestive of multiple pulmonary arterio-
venous fistulas, although the fistulas could not be visualized by angio-
cardiography. T h e child did not exhibit the other stigmata of multiple
hereditary telangiectasis.
Although it is stated that systemic manifestations of this hereditary
condition might exhibit themselves later in life, it is our contention
that multiple microscopic arteriovenous fistulas can be confined to
the lungs in the absence of visible aneurysms and in the absence of
systemic manifestations. These are probably the results of failure of
development of the pulmonary vascular capillaries. T h e term fistulas
should be reserved exclusively to this condition.

LARGE ARTERIOVENOUS ANEURYSM (PERIPHERAL)

This represents the most frequently encountered anomaly (Fig. 3).
The lesion usually consists of a single or a multiloculated sac. It is prob-
able that in addition to failure in development of the vascular capillaries
a coexisting failure is present in the primitive venous plexus, resulting
in persistence of a large venous channel of aneurysmal proportions. An
abnormal pulmonary artery branch supplies the aneurysm directly
and usually narrows abruptly before entry into the sac. Abnormal
contributions from neighboring vessels and lobes may also be present
[13]. T h e sudden narrowing prior to entry into the sac might represent
an attempt at separation in early embryonic life and might also be a
contributing factor to the aneurysmal enlargement distal to the stenotic
area. T h e venous drainage is usually a well-defined vascular channel
leading away from the aneurysm and emptying into the left atrium.
It is preferable to reserve the term aneurysm for this anomaly as a
distinction from the previously described condition.

LARGE ARTERIOVENOUS ANEURYSM (CENTRAL)

A centrally located aneurysm is a rare occurrence. We have recently
described a case [ Z ] in which the aneurysm was continuous with the
left atrium (Fig. 4). T h e aneurysm was located in the right lower lobe,
and its junction with the left atrium measured 5 to 6 cm. We are aware
of only one similar, unpublished case. This probably represents an
anomaly of the primitive large venous trunk posterior to the left atrium.
It is uncertain whether this has occurred during or following absorp-
tion of the vein into the left atrium.

VOL. 1, NO. 3, MAY, 1965 281

ANABTAWI, ELLISON, AND ELLISON

FIG. 3 . Pulmonary arteriogram showing the ,most common type of arterio-

venous aneurysm lying in the periphery of the lcft lung.

F I G . 4. Artist’s posterior view of an unusually large and centrally located

arteriovenous aneurysm of the right lower lobe encountered in an 8-year-old
male.

282 T H E ANNALS OF THOKACIC SURGERY

 Piilmonnry Arteriouenoiis Aneirrysrns and Fistiilas

LARGE AKTERIOVENOUS ANEURYSM WITH ANOMALOUS

VESOIIS 1)RAINAGE
111 1949 Grishman et al. [61 pointed out the presence of anomalous
vetioiis drainage in association with pulnionary arteriovenous aneu-
rpiis. ’l’hey presented three cases in which angiocardiography was
suggest i1-e of anomalous venous drainage. One patient underwent
surgery and the diagnosis was confirmed. T h e vein drained directly
into tlic inferior vena cava. T h e presence of cyanosis in all three
patictits is physiologically inconsistent with venous drainage of un-
oxygeiiated blood into the inferior vena cava. It is possible that venous
draiiiagc into the left atrium or multiple small arteriovenous fistulas
were o\w-looked. Therefore, it is well to emphasize the need for angio-
cardiography and verification of the exact anatomical abnormality at
surgcry. T h e cast technique used by Lindskog and colleagues [I 31
might serve as a useful adjunct to the study of their structure. Recently
Ablmtt and Leigh [I] described a case of pulmonary arteriovenous
aneiirysiii of the left upper lobe with aberrant venous drainage to the
left ititiominate vein. This partially imitated the left cardinal vein.
T l q stated that the same situation was encountered in two other
pit iciits--one in association with persistent ductus arteriosus and the
other \\.it11 interatrial septal defect and transposition of the right pul-
nioii;iry ipeins.

XIIII.l‘II’1.I.: SMALL ARTERIOVENOUS FISTULAS WITH

ASO>IAI.OUS V E N O U S DRAINAGE
‘Ilw mechanism of hypoxia in association with left-to-right shunts,
particularly atrial septal defects, has been investigated by several
workers. Explanations given include: streaming of blood from the
inferior vena cava and right atrium, large blood flow across the pul-
monary bed, and the presence of pulmonary arteriovenous shunts. T h e
great capacity of the lung to oxygenate large quantities of blood and
persistence of venous admixture following corrective surgery in some
cases suggests the presence of pulmonary arteriovenous fistulas [ 5 ] .
This is supported by the presence of mild to moderate desaturation in
preoperative blood samples taken from the pulmonary veins in the
course of cardiac catheterization (Table 1). Whether this persists follow-
ing closure of the atrial septal defect cannot be determined.

LARGE V E N O U S ANEURYSM WITHOUT FISTULAS

It is quite conceivable that abnormal development in the veins
peripherally might occur in the absence of pulmonary fistulas. Cases
of pulmonary aneurysms have been reported in the absence of fistulas
or cyanosis. Systemic artery communication was present in some b u t
not all cases. Watson [24] reported a case in which the aneurysm was
supplied by an aberrant artery from the aorta. T h e patient was asympto-

VOI.. I , NO. 3, MAY, 1965 283

ANABTAWI, ELLISON, AND ELLISON

TABLE 1. OXYGEN SATURATION I N P A T I E N T S WITH

ANOMALOUS VENOUS DRAINAGE

Peripheral Artery Pulmonary Vein

Patient Hgb. O2 Content Hgb. O2 Content
No. Diagnosis Sat. O2 Vol. % Sat. 0, Vol. %
1 LV to RA shunt 96 16.05 90 15.26
2 ASD 97 19.18 95 18.70
3 ASD 94 18.79 84 16.93
4 ASD 95 15.94 87 14.61
5 ASD 90 16.54 89 16.40
6 ASD 97 2 1.76 87 19.48

matic and acyanotic. Following ligation of the feeder artery the

operator was unable to palpate the cystic mass in the lung. The mass
was not resected. Prutzman and Flick [15] reported aberrant supply
from the intercostal arteries. Absence of cyanosis does not necessarily
mean that fistulas are not present, as shown by Lindskog et al. [13] using
Vinylite casts of resected specimens. Well-developed fistulas were
demonstrated in the absence of cyanosis. No satisfactory explanation
was given, but it is quite possible that the shunt was small or anomalous
venous drainage was present.
T h e presence of systemic communications with the aneurysm is
not surprising. Precapillary systemic-pulmonary anastomoses have been
demonstrated in normal [23] and abnormal states such as bronchiectasis
and congenital heart disease [lo, 11, 221. It is conceivable that ab-
normally large bronchial communications may be the sole arterial
contribution to the aneurysm.

ANOMALOUS VENOUS DRAINAGE

Discussion of this anomaly does not lie within the scope of this
paper. It is listed only for the purpose of completing our proposed
classification.

SUMMARY
Anatomical and clinical variations of the pulmonary vascular
system have been discussed. Most of these anomalies are explainable
in the light of present knowledge of the embryology of this system.
T h e relationship of fistulas and aneurysms to anomalous venous drain-
age has also been emphasized, and a classification for them has been
suggested.

REFERENCES
1. Abbott, 0. A., and Leigh, T. F. Aneurysmal.dilatations of the superior
vena caval system. A n n . Surg. 159:858, 1964.

284 THE ANNALS OF THORACIC SURGERY

 Pulmonary Arteriovenous Aneurysms and Fistulas

2. Anabtawi, I. N., and Ellison, K. G. Maldevelopment of pulmonary veins

and pulmonary arteriovenous aneurysms. Amer. Surg. 30: 770, 1964.
3. Brink, A. J. Telangiectasis of lung. Quart. J. Med. 19:239, 1951.
4. Cooley, D. A., and McNamara, D. G. Pulmonary telangiectasis: A report
of a case proved by pulmonary biopsy. J. Thorac. Surg. 27:614, 1954.
5. Ellison, L. T., Hall, D. P., Yeh, T . , Mobarhan, H., Rossi, J., and Ellison,
R. G. Physiological alterations in increased pulmonary blood flow with
and without pulmonary hypertension. J. A p p l . Physiol. 16:305, 1961.
6. Grishman, A., Poppell, M. H., Simpson, R . S., and Sussman, M. L. T h e
roentgenographic and angiocardiographic aspects of (1) abnormal insertion
of pulmonary veins associated with interatrial septa1 defect and (2) con-
genital arteriovenous aneurysm of the lung. Amer. ]. Roentgen. 62500, 1949.
7. Hepburn, J., and Dauphinee, J. A. Successful removal of hemangioma
of the lung followed by disappearance of polycythemia. Amer. J. M e d . Sci.
204:681, 1942.
8. Jose, A. B., and Milnor, W. R. T h e demonstration of pulmonary arterio-
venous shunts in normal human subjects and their increase in certain
disease states. J. Clin. Invest. 38: 1915, 1959.
9. Le Roux, B. T. Pulmonary arteriovenous fistulae. Quart. J. M e d . 28: 1, 1959.
10. Liebow, A. A., Bloomer, W. E., Harrison, W., and Lindskog, G. E. Studies
on the lung after ligation of the pulmonary artery: 11. Anatomical changes.
Amer. J . Path. 26:177, 1950.
11. Liebow, A. A., Hales, M. R., and Lindskog, G. E. Enlargement of the
bronchial arteries and their anastomoses with the pulmonary arteries in
bronchiectasis. Amer. J. Path. 25:211, 1949.
12. Lilienthal, J. L., Jr., Riley, R . L., Proemmel, D. D., and Franke, R . E. An
experimental analysis in man of the oxygen pressure gradient from alveolar
air to arterial blood during rest and exercise at sea level and altitude. Amer.
J. Physiol. 147: 199, 1946.
13. Lindskog, G. E., Liebow, A. A., Kansel, H., and Janzen, A. Pulmonary
arteriovenous aneurysm. Ann. Surg. 132:591, 1950.
14. Prinzmetal, M., Ornitz, E. M., Jr., Simkin, B., and Bergman, H. C. Arterio-
venous anastomoses in liver, spleen and lungs. Amer. J. Physiol. 152:48,
1948.
15. Prutzman, L. D., and Flick, J. B. Pulmonary arteriovenous fistula with
extensive thoracic wall collateral circulation. Bull. Aye7 Clin. Lab. 4:23, 1954.
16. Rodes, C. B. Cavernous hemangioma of the lung with secondary poly-
cythemia. J.A.M.A. 110: 1914, 1938.
17. Shenstone, N. S. Experiences with total pneumonectomy. J. Thorac. Surg.
11:405, 1942.
18. Sloan, R. D., and Cooley, K. N. Congenital pulmonary arteriovenous
aneurysm. Amer. J. Roentgen. 70: 183, 1953.
19. Steinberg, I., and McLenahan, J. L. Pulmonary arteriovenous fistula:
Angiocardiographic observations in nine cases. Amer. J. M e d . 19:549, 1955.
20. Tobin, C. E. T h e bronchial arteries and their connections with other
vessels in the human lung. Surg. Gynec. Obstet. 95:741, 1952.
21. Tobin, C. E., and Zariquiey, M. 0. Arteriovenous shunts in the human
lung. Proc. SOC. Exp. Biol. (N. Y.) 75:827, 1950.
22. Turner-Warwick, M. Precapillary systemic pulmonary anastomoses. Thorax
18:225, 1963.
23. Verloop, M. C. T h e bronchial arteries and their anastomoses with the
arteriae pulmonales in the human lung: A microanatomical study. Acta
Anat. (Basel) 5:171, 1948.
24. Watson, w. L. Pulmonary arteriovenous aneurysm: A new surgical disease.
Surgery 22:919, 1947.
25. Whiddon, M., Allen, L., and Ellison, R. G. Unpublished observations, 1964.

VOL. 1 , NO. 3, MAY, 1965 285