U.S. flag

An official website of the United States government

Skip to main content
Return to topReturn to top



Arterial Aneurysms and Arteriovenous Fistulas

Successful Suture of the Abdominal Aorta for Arteriovenous Fistula

Norman E. Freeman, M. D. and Ambrose H. Storck, M. D.

   An arteriovenous fistula involving the abdominal aorta is seldom observed since death usually follows promptly from the massive hemorrhage caused by the original injury. Even if the patient recovers from the initial loss of blood, as in the two cases reported by Makins1 in which the patients survived for several weeks, the short-circuiting of the circulation immediately places such a strain on the heart that early cardiac failure occurs.

   In 1944, Pemberton, Seefeld, and Barker 2 successfully repaired an arteriovenous fistula occurring between the abdominal aorta and the inferior vena cava.   They were unable, in a comprehensive review of the literature, to find any previously reported case in which the patient had survived for a sufficiently long period of time after surgical repair to furnish evidence that the operation had been successful. The case described in this chapter is apparently the second to be recorded in which a successful repair of the abdominal aorta has been performed.


   The patient, a 25-year-old infantryman, was wounded on Okinawa 14 May 1945. A bullet from a .25-caliber rifle entered the abdomen 3 inches below the ensiform cartilage just to the right of the midline and passed out through the back at the level of the second lumbar vertebra. He immediately became paralyzed below the waist. A laparotomy was performed the day of injury and a large retroperitoneal hematoma found. The abdomen was closed without drainage. The patient stated that he was told that numerous veins had been tied off during the operation. Ten days after injury he was transferred to another hospital where roentgenograms were made. They disclosed several fracture lines in the spinous process of the second lumbar vertebra radiating through the lamina without displacement or separation. On the 15th day after injury, a laminectomy was performed and the comminuted fragments of the spine at the first and second lumbar vertebrae, and the lamina of the second lumbar vertebra, removed. The underlying dura was found to be compressed and lacerated and three of the nerve roots severed.

1 Makins, G.H.: On Gunshot Injuries tothe Blood Vessels, Founded on Experience Gained inFrance During the Great War, 1914-1918. Bristol, John Wright &Sons, Ltd., 1919.

2 Pemberton, J. deJ.; Seefeld, P. H., and Barker, N. W.:Traumatic arteriovenous fistula involvingthe abdominal aorta and the inferior vena cava. Ann. Surg. 123:580-590, Apr 1946.


    After this operation there was considerable return of function in the lower extremities. On the fifth postoperative day, swelling of the right leg appeared and a diagnosis of thrombophlebitis was made.  This swelling subsided in a few days.

    Preoperative Clinical Course. Six weeks after his originalinjury the patient complained of someepigastric pain and, on examination, a pulsating mass with intensethrill was found in the upperabdomen. A diagnosis of aneurysm of the abdominal aorta was made and hewas evacuated to theZone of Interior.

    When admitted to the debarkation hospital his blood pressure was 150 mm. of mercury systolic and 60 diastolic. Hemaglobin was 72 percent of normal and urine examinations showed some infection to be present. Tidal drainage was instituted to correct this condition. Some dilated veins over the abdominal wall were observed. Physical examination revealed nothing which would indicate cardiac enlargement. That the heart was of normal size was confirmed by roentgenogram. A chemical examination of the blood revealed it to be essentially normal, and the result of the Kahn test was negative.

    The patient was transferred to the vascular center at DeWitt General Hospital 3 August 1945. Examination at this time revealed a pulsating mass in the epigastrium which was more prominent on the right of the midline. The superficial abdominal veins were dilated (Fig. 44).

    Acontinuous loud bruit, which was accentuated during systole, could beheard over the mass in theabdomen. The pulse rate was 96 beats per minute and the blood pressure152 mm. of mercury systolicand 96 diastolic. The heart was not enlarged and there was no 

Figure 44. Infrared photograph of patient with arteriovenous fistula involving abdominal aorta and vena cava. Note dilatation of veins of abdomen and thorax.


evidence of dilatation or engorgement of the neck veins. The patient was neitherdyspneic nororthopneic. Venous pressure measured in the right antecubital vein was3 cm. of water. The lungswere clear and the liver edge palpable at the right costal margin.????

Neurologic examination revealed a residual paraplegia involving principally the motor components of the posteriort tibial and common peroneal nerves on the right side, and some involvement of the peroneal nerve on the left side. The sensory loss was small and was confined to a small area about the anus and scrotum on the right side.

     It was not possible to obliterate the thrill by pressure in the epigastrium or right upper quadrant. The fact that the heart showed no evidence of marked strain, even 3 months after the development of the lesion, was interpreted as evidence that there was some interference with the free return of blood from the arterial to the venous side of the circulation. Since the liver was not enlarged, it was felt that there was no involvement of the portal venous system. The dilatation of the superficial abdominal veins suggested involvement of the inferior vena cava. A diagnosis was made of arteriovenous fistula probably involving a branch of the abdominal aorta and either the vena cava or some tributary of this vein.<>

    The patient was placed on the paraplegic ward where he was encouraged to increase his activities as much as was compatible with his residual neurologic lesions. During his stay the bladder infection was controlled by tidal drainage, his appetite improved, and he regained some of the weight which he had lost.

    On 15 September, 4 months after the initial injury, the patient complained of some abdominal discomfort and vomited. Examination at this time disclosed the abdominal mass to be approximately 6 cm. in diameter. Both the pulsation and the thrill were more apparent than they had been. Gastrointestinal examination with a barium meal failed to reveal evidence of any extrinsic mass producing pressure upon the pylorus or duodenum. The cardiac consultant noted a definite increase in the size of the liver with an increase in the pulse rate and expressed the opinion that the patient was showing evidence of cardiac strain. Because of the possibility of cardiac damage and the increase in the size of the mass it was decided to operate without further delay.<>

. The operation was carried out with the patient under intratracheal ether oxygen anesthesia. A right paramedian incision was made from the xyphoid to just below the umbilicus. Numerous dilated veins were encountered in the subcutaneous tissues. When the peritoneum was opened, a large pulsating mass was found beneath the gastrohepatic omentum. A puckered scar was present near the border of the liver on the right side and probably represented the point of entrance of the rifle bullet. The aneurysm, which lay behind the vena cava, displaced the vena cava forward and so compressed it as to hinder the ready flow of blood back to the right side of the heart. The veins of the portal system did not appear to be dilated. The hepatic, common, and cystic ducts were readily visualized and appeared to be pushed forward by the pulsating mass which occupied the posterior aspect of the right upper quadrant. The aneurysm was under considerable pressure and at one point, below and slightly medial to the gallbladder, the thrill of the arteriovenous fistula was most easily palpable. While compression at this point obliterated the thrill, it also appeared to produce an increase in the intea-aneurysmal pressure<>.

    The round ligament of the liver and some adhesions were divided. An attempt was made to visualize the artery entering the aneurysm by dividing the gastrohepatic omentum. This, however, still did not permit localization of the opening of the artery into the aneurysm. Only by pressure on the aorta at the hiatus of the diaphragm was it possible both to obliterate the thrill and to cause the aneurysmal sac to collapse. The aorta was therefore exposed at this point by dividing some of the fibers of the diaphragm. It was then encircled by a fine rubber catheter fitted to a Bethune tourniquet. Attempts to expose the aorta through the root of the mesentery beneath the transverse colon were unsuccessful because of the dilated veins in this region.<>

    The perioteium and transversalis fascia were next incised from within the abdomen just to the left of the midline. By separating these structures from the undrlying muscles


it was possible to expose the anterior surface of the psoas muscle and the vertebral column. The spleen, descending colon, pancreas, left kidney, and intestines were reflected to the right and the abdominal aorta exposed as it lay on the anterior surface of the lumbar vertebrae. Many large veins were divided and ligated. The tissues about the aorta were thickened and edematous. The discoloration which was present indicated old hemorrhage. The abdominal aorta was encircled by a segment of rubber tubing just proximal to the origin of the inferior mesenteric artery. It was then exposed just above the origin of the left renal artery where it was again encircled by a piece of tubing. Compression of the aorta by this piece of tubing caused the sac to collapse. An additional section of tubing was placed about the left renal artery. With the proximal and distal aorta and the left renal artery occluded, the inflammatory tissues surrounding the aorta at the site of the fistula were incised and the aorta was finally cut away from the aneurysm at this location. Figure 45 illustrates the location of the lesion. <>

The opening into the aorta measured one-half inch in length.  It was closed by a transverse running stitch of No. 0000 Deknatel which had been passed through sterile mineral oil.  Bleeding from the sac was only moderate and was readily controlled by digital pressure. No effort was made to excise the sac.  This opening was closed by a running stitch of No. 0000 silk. At the conclusion of this procedure, the segments of tubing around the distal aorta, left renal artery, and proximal aorta were released in that order. Good pulsation, expansile in character, was apparent in the aorta below the suture line. No bleeding took place. Two Penrose drains were inserted through a stab wound below the left costal margin into the region of the left  lumbar gutter and the abdomen was closed.

    During the operation, which lasted almost 8 hours, the patient received a continuous transfusion of 3,000 cc. of whole blood and 500 cc. of physiologic salt solution. When the abdominal aorta was occluded the blood pressure increased from 118 mm. of mercury systolic and 70 diastolic to 240 systolic and 100 diastolic. The pulse rate rose from 130 to 160 beats per minute and the neck veins became greatly distended. When the tourniquet, which had been in place for 1 hour and 40 minutes was released, the systolic pressure fell to 50 mm. of mercury. Within 30 minutes, however, it rose to 110 mm. of mercury. The diastolic pressure at this time was 80 and the pulse rate was 140 beats per minute.

    Postoperative Course. Immediately after operation a strong femoral pulse was palpable and within an hour the pulse at the wrist was of good volume and all extremities warm and dry. The patient was placed in an oxygen tent and continuous intestinal decompression therapy was instituted by means of suction applied to an indwelling Levin tube. By the following morning the abdomen was flat and peristalsis present. The patient was conscious and alert. Because of persistent low blood pressure and a rapid, weak pulse, he was given another transfusion of whole blood. Blood pressure then rose to 160 mm. of mercury systolic and 90 diastolic. Administration of penicillin and sulfadiazine was begun, but owing to urinary suppression the sulfadiazine was discontinued after the administration of 7 cm. in the first 36 hours.

    Impaired renal function presented a serious complication. During the first 48 hours, in spite of receiving 3,500 cc. of 5-percent glucose in distilled water and 1,000 cc. of 5-percent glucose in physiologic salt solution, the patient voided only 200 cc. of urine. Elevation of the nonprotein nitrogen following operation is shown in Table 34.

    For the first week after operation treatment consisted chiefly of continuous oxygen therapy, decompression of the upper gastrointestinal tract, and the administration of fluids by vein in amounts just sufficient to balance the losses through the gastrointestinal tract and kidneys, and by insensible loss of water.

    Three days after operation the venous pressure in the right antecubital vein was 12.6 cm. of saline solution. The patient was feeling well, but he had no appetite and his mouth was sore because of superficial erosions of the mouth and lips. Blood pressure was consistently 170 mm. of mercury systolic and 80 diastolic.


Figure 45. Diagrammatic representation of findings at operation for arteriovenous fistula of abdominal aorta and vena cava.




    Six days after operation, ophthalmologic examination was reported as showing "remarkable generalized narrowing of the retinal arterioles throughout all divisions. In many of the vessels there are variations in caliber indicative of focal spasm. No signs of sclerosis are noted, no hemorrhages or edema." A diagnosis of acute retinal angiospasm was made.

    Because of his sore mouth and lack of appetite the patient refused to eat. On the 10th postoperative day, therefore, high caloric feedings were started. In 36 hours the patient received 1,500 cc. of fluid containing approximately 1,800 calories. The feedings were given by continuous drip through a nasal tube. He tolerated this feeding very well and his general condition improved rapidly. The nonprotein nitrogen of the blood dropped as the volume of urinary excretion increased. The wound healed without complications.

    The day after operation the patient noted that he could not dorsiflex the left foot. Neurologic examination showed a definite increase in the neurologic disturbances noted prior to operation. In addition to the foot drop on the left side, the area of anesthesia had in creased and there was a decrease in the bladder tone.  For the first 2 months after operation severe burning pain was experienced in both feet, but this condition suddenly cleared up at the end of this period with a concurrent improvement in motor power.

    Evidence of renal damage persisted for several weeks with a constantly low specific gravity of the urine and persistent mild hypertension, and even 6 weeks after operation the ophthalmologic examination revealed moderate generalized narrowing and increased tortuosity of the retinal vessels. Excretory urograms made 6 weeks after operation revealed the excretory function of the kidneys to be normal. The final urine concentration test, made 2 months after operation, showed an ability of the kidneys to concentrate the urine to 1.018. Roentgenograms, made at this time, showed a defect in the laminae between the second and third lumbar vertebrae at the site of the original fracture.

     About 2 ½ months after operation the patient was transferred to another general hospital. At this time he had recovered sufficiently from the spinal cord injury to walk with the aid of one cane.

    Eight months after operation it was reported that the patient had shown no evidence of recurrence of the fistula.


    The presence of a fistula between the abdominal aorta and vena cava usually leads to rapid heart failure and death. The absence of this complication in this case can probably be explained by the intervention of a large


aneurysmal sac between the two vessels. The dilatation of the superficial abdominal veins observed before operation is in keeping with this explanation. The transient swelling of the right leg noted 10 days after injury (which was originally diagnosed as thrombophlebitis) was probably the result of interference with the return flow of blood from the lower extremity.

    The increase in the paralysis of the bladder and lower extremities following operation was not surprising in view of the fact that the abdominal aorta was completely occluded for 100 minutes. After complete occlusion of the aorta for a period of 40 to 55 minutes, Blalock and Park 3 observed paralysis of the hind quarters in the dogs used in their experiments. The neural damage evident after operation might be attributed to the result of impairment of the circulation to the spinal cord and cauda equina or to the temporary ischemia of the distal nerves, but the involvement of the nerves to the bladder suggest the former explanation. The rapid improvement which was observed 2 months after operation indicated a favorable prognosis.4

    The second postoperative complication was the temporary urinary suppression, associated with hypertension. This complication was probably the result of the renal ischemia produced by occlusion of the abdominal aorta above the renal arteries. It was associated with marked vasospastic changes in the eyegrounds and with nitrogen retention. With resumption of renal function at the end of 2 weeks, the hypertension subsided. The final urine concentration test, which was done 2 months after operation, showed an ability to concentrate to 1.018. The excretory urogram was also quite satisfactory, but it is possible that some permanent damage to the kidneys was sustained. In a case reported by Alexander and Byron 5 in which a segment of the thoracic aorta was resected, hypertension with severe retinal angiospasm, exudates, and hemorrhages proved a serious late complication. In a case reported by Pemberton, Seefeld, and Barker, 6 persistent hypertension with cardiac hypertrophy was also noted.

    During the operation considerable difficulty was experienced in locating the opening of the artery into the aneurysmal sac. Reflection of the duodenum with exposure of the anterior surfaces of the aorta and vena cava was employed by Pemberton and his associates 7 as the method of exposure, but the aneurysmal sac in their patient lay to the left of the aorta. Since in this case the sac appeared to lie between the aorta and the vena cava, a similar approach could not be used. It was only after retroperitoneal exposure of the anterior surface of the psoas muscle and the lumbar vertebrae by displacement of the abdominal

3 Blalock, A. and Park, E.A.: Surgical treatment of experimental coarctation (atresia) of the aorta. Ann. Surg. 119: 445-456, Mar 1944.
4 When this patient was examined a year after the operation at the time of his discharge from the Veterans Hospital at Van Nuys, Calif., he could walk with the aid of a single cane and could go up and downstairs. There was no recurrence of the abdominal aneurysm or arteriovenous fistula. Bladder control was incomplete, but he could retain 12 oz. of urine. Subsequently a letter was received from the patient saying that he had returned to work (cabinetmaking) and was able to hold a full-time job.
5 Alexander, J., and Byron, F. X.: Aortectomy for thoracic aneurysm. J. A. M. A. 126: 1139-1145, 30 Dec 44.
6 See footnote 2, p.302.



contents from the left lumbar gutter that the abdominal aorta could readily be exposed. It was then possible to visualize the entire length of this vessel from the diaphragm to its bifurcation.


Transvenous suture of the opening between the aorta and vena cava, the technique known as the Matas-Bickham 8 operation, was used by Pemberton and his associates. Although it is frequently valuable, it has the disadvantage of not permitting inspection of the entire arterial wall and arterial aneurysms have been known to develop after its use in cases in which there happened to be additional weakened points in the arterial wall. Closure of an arteriovenous fistula leads to a marked increase in blood pressure within the artery at the site of the fistula.9 If the wall close to the former communication is defective it may give way and result in the formation of an aneurysm. On the other hand, complete dissection of the artery from the fistula permits thorough examination so that other damaged portions of the wall are unlikely to escape notice. Excision of the damaged portion of the arterial wall with transverse closure of the defect is the procedure of choice under these circumstances. 10

8 Matas, R.: Treatment of arteriovenous aneurisms by intrasaccular method of suture (endo-aneurismorrhaphy) with special reference to transvenous route. Ann. Surg. 71: 403-427, Apr 1920.
9 Freeman, N. E.: Direct measurement of blood pressure within arterial aneurysms and arteriovenous fistulas. Surgery 21: 646-658, May 1947.

10Freeman, N. E.: Arterial repair in the treatment of aneurysmsand arteriovenous fistulae; reportof 18 successful restorations. Ann. Surg. 124: 888-919, Nov 1946.