Endovascular Grafts and Other Image-Guided Catheter-Based Adjuncts to Improve the Treatment of Ruptured Aortoiliac Aneurysms

Abstract

Objective

To report a new management approach for the treatment of ruptured aortoiliac aneurysms.

Methods

This approach includes hypotensive hemostasis, minimizing fluid resuscitation, and allowing the systolic blood pressure to fall to 50 mmHg. Under local anesthesia, a transbrachial guidewire was placed under fluoroscopic control in the supraceliac aorta. A 40-mm balloon catheter was inserted over this guidewire and inflated only if the blood pressure was less than 50 mmHg, before or after the induction of anesthesia. Fluoroscopic angiography was used to determine the suitability for endovascular graft repair. When possible, a prepared, “one-size-fits-most” endovascular aortounifemoral stented PTFE graft was used, combined with occlusion of the contralateral common iliac artery and femorofemoral bypass. If the patient’s anatomy was unsuitable for endovascular graft repair, standard open repair was performed using proximal balloon control as needed.

Results

Twenty-five patients with ruptured aortoiliac aneurysms (18 aortic, 7 iliac) were managed using this approach. Balloon inflation for proximal control was required in nine of the 25 patients. Twenty patients were treated with endovascular grafts. Five patients required open repair. The ruptured aneurysm was excluded in all 25 patients; 23 survived. Two deaths occurred in patients who received endovascular grafts with serious comorbidities. The surviving patients who received endovascular grafts had a median hospital stay of 6 days, and the preoperative symptoms resolved in all patients.

Conclusions

Hypotensive hemostasis is usually an effective means to provide time for balloon placement and often for endovascular graft insertion. With appropriate preparation and planning, many if not most patients with ruptured aneurysms can be treated by endovascular grafts. Proximal balloon control is not required often but may, when needed, be an invaluable adjunct to both endovascular graft and open repairs. The use of endovascular grafts and this approach using other image-guided catheter-based adjuncts appear to improve treatment outcomes for patients with ruptured aortoiliac aneurysms.

Four decades have passed since the first surgical repair of a ruptured infrarenal abdominal aortic aneurysm (AAA) was reported by Gerbode in 1954. ¹ During this period, several important advances have been made in the nonsurgical aspects of care. These include improvements in the transportation of these patients, critical care, and management of cardiac dysfunction and pharmacologic support. Despite these efforts, surgical death rates have not improved significantly and still range from 24% to 70%. ^2–22 As a result, most of the recent literature on ruptured AAAs has focused on selecting patients who are likely to survive, and some authors have advocated abandoning treatment for certain groups of patients who have risk factors that predict a poor outcome. 5,6,8,9,11,15,16,19,20,22 In addition, it has been thought that further improvement in surgical outcome is unlikely; therefore, the focus has shifted toward screening patients to treat aneurysms before they rupture 3,8,12,18–20.

We believe the reason why the outcome after surgical repair of ruptured aortoiliac aneurysms (AIAs) has not improved relates to the fact that the basic surgical techniques for repairing ruptured AIAs remain little changed in the past four decades, although several improvements have been introduced. These improvements include liberal use of supraceliac clamping, suturing the graft within the aneurysm without excising it, more frequent use of tube grafts, and use of lower-porosity prostheses. 3,4,6,7,9–11,13 However, general anesthesia (which often exacerbates hypotension by releasing the sympathetic vasoconstriction), abuse of fluid resuscitation, a large surgical incision and dissection, and the accompanying invasiveness and blood loss required for open repair have largely remained unchanged and may be responsible for the poor outcome.

Endovascular grafts (EVGs) have been used to treat a variety of arterial pathology, including elective AAAs, iliac artery aneurysms, occlusive disease, and traumatic lesions, with promising short-term and midterm results. ^23–32 Despite the theoretical advantage of EVGs for the treatment of ruptured AAAs, their use has been limited to one small series and several case reports. ^33–37 One major reason for this limited use of EVGs in this acute setting has been the need for preoperative measurements of the aneurysmal and adjacent arterial anatomy so that the appropriate size and configuration of graft can be selected. The resulting delay has been deemed inappropriate in the urgent setting of ruptured AAAs.

We previously reported our initial results on the feasibility of EVG repair for ruptured AAAs. ³⁴ In that report we described our method of facilitating EVG repair of ruptured AIAs, which features the use of a “one-size-fits-most” EVG that can be customized during surgery to avoid delay in the fabrication and procurement of the graft, and selective use of a transbrachially deployed supraceliac balloon before EVG insertion for proximal control. The purpose of this paper is to report the advances we have made in the EVG treatment of ruptured AIAs. We will also describe how other fluoroscopic-guided, catheter-based techniques can determine which patients are suitable for EVGs and which will require open repair, and we will show how these endovascular techniques can facilitate rapid and minimally invasive proximal balloon control, which can be helpful in both forms of repair. With these techniques, the approach to management of ruptured AIAs may change in a way that improves the treatment outcome.

METHODS

This study started on April 21, 1994, when we performed the first EVG repair of a ruptured AIA. ²⁵ Since then, we have continued to improve and extend the use of our EVG and protocol for the treatment of AIAs ³⁴ (Figs. 1 through 4). The patients and methods presented in this paper represent an evolutionary process. Because the initial methods were described in our earlier report, the present article will focus on the methods used in our current protocol.

Figure 1. The main graft and the occluder device of the Montefiore Endovascular Grafting System. G, gold bead denoting cranial end of expanded polytetrafluoroethylene graft; P, Palmaz stent; L, ligature.

Figure 2. An endovascular graft (EVG) repair of a complex aneurysm repair using the Montefiore Endovascular Grafting System. The cranial end of the graft is fixed within the proximal neck with a large Palmaz stent (S). In this example, the bare portion of the stent is deployed across the orifice of the renal arteries. The distal end of the EVG is secured to the femoral artery by a hand-sewn endoluminal anastomosis (E). The occluder device (O) is deployed in the contralateral common iliac artery to preserve at least one internal iliac artery. C, embolization coils; F, femorofemoral bypass.

Figure 3. Tailoring the proximal stent to varying neck diameters by changing the inflation pressure of the balloon. (A) The proximal stent is expanded to 20 mm when the deployment balloon is inflated to 2 atm. (B) At 6 atm of inflation pressure, the stent is expanded to 28 mm.

Figure 4. Tailoring the length of the endovascular graft (EVG). (A) The EVG is always made long enough so that its distal end (D) emerges from the femoral arteriotomy site. (B) The distal end of the EVG is cut to appropriate length and a hand-sewn endoluminal anastomosis (E) is performed within the femoral artery. This arteriotomy site is used as the anastomosis site for the following femorofemoral bypass.

Patients

During the past 6 years, 25 patients with ruptured AIAs were candidates for treatment with an EVG (Table 1). All 25 patients had an acute onset of symptoms, including severe abdominal or back pain (n = 21), syncope (n = 8), or external bleeding (n = 4). Twenty (80%) were treated with an EVG; five (20%) required an open repair. Twenty-five percent of the patients treated with an EVG and 40% of those treated in a standard open fashion had hypotension (systolic blood pressure <90 mmHg at any time before surgery). There were 18 AAAs and 7 isolated iliac artery aneurysms (Table 2).

Table 1. PATIENT DEMOGRAPHICS

COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; EVG, endovascular graft; SBP, systolic blood pressure.

Table 2. ANEURYSM CHARACTERISTICS

EVG, endovascular graft.

Initially we used EVGs only in hemodynamically stable patients, but as we gained experience we broadened our indications and have tried to use EVGs whenever possible. For the past 18 months, all 15 patients diagnosed with a ruptured AIA, including those in profound shock, were entered into a protocol in which they were taken to the operating room, a brachial wire was placed under local anesthesia, and an arteriogram was performed. Based on the angiographic findings, a decision was made to proceed with EVG or open repair. Ten (67%) of the 15 patients were treated with an EVG; 5 (33%) required an open repair.

Preoperative Management

Hypotensive Hemostasis

As advocated by Crawford, ⁷ we minimized the use of fluid resuscitation before obtaining the means for aortic control with a brachially introduced balloon, as described below. Limited fluid resuscitation was performed only when the systolic blood pressure fell to less than 50 mmHg. Fluid resuscitation was withheld as soon as the blood pressure exceeded 50 mmHg.

Immediate Transfer to the Operating Room

When the diagnosis of presumed ruptured AIA was made, the patient was taken to the operating room. ³⁸ In hemodynamically stable patients or in those in whom the diagnosis was questionable, a preoperative spiral computed tomography scan was performed with or without intravenous contrast.

Routine Brachial Guidewire Placement Before General Anesthesia Induction

As soon as the patient was brought to the operating room, a 5F sheath was inserted in the right brachial artery (antecubital fossa) under local anesthesia (Fig. 5). Then, using directional catheters, a Glidewire (0.035″, 260 cm; Meditech, Oakland, NJ) was placed in the descending thoracic aorta under fluoroscopic guidance (Model 9800, OEC, Salt Lake City, UT, or BV 300 Plus, Philips, Eindhoven, The Netherlands). In patients who were hemodynamically unstable (blood pressure <50 mmHg), the 5F sheath was exchanged for a 14F sheath (Cook, Bloomington, IL) and an occlusion balloon catheter (40-mm occlusion catheter, Meditech) was inserted over the guidewire and inflated in the supraceliac aorta (Fig. 6A). In addition, this proximal control was obtained with the occlusion balloon in patients whose blood pressure fell to less than 50 mmHg after general anesthesia induction led to the release of vasoconstriction that had maintained the blood pressure at more than 50 mmHg.

Figure 5. The endovascular operating room equipped with a portable digital fluoroscope (F). Note the insertion of the brachial wire (B) before induction.

Figure 6. A 64-year-old man was found unconscious at home. On arrival in the emergency room, his systolic blood pressure was 40 mmHg, the abdomen was distended, and there was a pulsatile mass. A diagnosis of presumed ruptured aortoiliac aneurysm (AIA) was made, and he was taken to the operating room without obtaining a preoperative computed tomography scan. (A) A brachial wire (B) and subsequently an occlusion balloon (O) were placed in the abdominal aorta under local anesthesia. On inflating the occlusion balloon in the supraceliac aorta, the systolic blood pressure immediately rose to 110 mmHg. Note the curled brachial wire, suggestive of a large AIA. (B) A diagnostic aortogram was obtained with the occlusion balloon inflated. The aortogram confirmed the presence (diagnosis) of an abdominal aortic aneurysm with a well-defined infrarenal neck, fulfilling the inclusion criteria for endovascular graft repair. Note the underfilled visceral and renal arteries resulting from proximal occlusion and lack of prograde flow. (C) After the femoral arteries were exposed and the sheath was inserted into the abdominal aorta, the supraceliac occlusion balloon (O) was deflated and exchanged for an infrarenal occlusion balloon (I) inserted from the femoral artery. (D) An aortogram was repeated by means of the brachial catheter to confirm perfusion of the visceral and renal arteries. (E) Completion angiogram revealed complete exclusion of the large aneurysm with no evidence of an endoleak. P, proximal Palmaz stent. (F) Aneurysm-sac-gram was performed by means of a catheter inserted from the left femoral artery after the deployment of the main graft system. The intrasac pressure measured through the catheter was 30 mmHg; the radial artery pressure was 120 mmHg. The contrast remained in the sac after injection, suggesting lack of an endoleak. (G, H, I) Postoperative contrast-enhanced computed tomography scan. The aneurysm is completely excluded from the circulation and the contrast is confined within the endovascular graft (E). Although this patient underwent evacuation of the hematoma for abdominal compartment syndrome, a considerable amount of hematoma (H) remained in the retroperitoneal space. (J) A limited transabdominal incision was made to evacuate the large hematoma. Five thousand milliliters of blood and clot were evacuated from both the intraperitoneal and retroperitoneal spaces, confirming overt rupture into the peritoneal cavity. The retroperitoneum was opened, and the anterior wall of the aneurysm (A) was exposed and the lack of hemorrhage was confirmed.

On-Table Angiogram and Inclusion Criteria for EVG Repair

An angiogram was performed through a brachial catheter using the Acist contrast power injector system (Acist Medical Systems, Eden Prairie, MN) (see Fig. 6B). If proximal balloon inflation was required, the arteriogram could be performed through the guidewire channel. A decision was made regarding the feasibility of an EVG repair, based on several inclusion criteria: the infrarenal neck length had to be longer than 10 mm; the proximal neck diameter had to be less than 28 mm; there could be no long bilateral iliac artery occlusions; and there could be no evidence of mycotic aneurysm. If the patient met these criteria, an EVG was used. Standard open repair was reserved for those who did not meet the EVG inclusion criteria.

EVG Repair

Under general anesthesia, one or both common femoral arteries were exposed and an introducer sheath was inserted in the aorta. If the supraceliac proximal occlusion balloon had been used to stabilize the blood pressure, an infrarenal occlusion balloon was introduced and deployed under fluoroscopic control, and the supraceliac balloon was then deflated (see Fig. 6C). An aortogram was repeated to confirm the perfusion of the visceral and renal arteries (see Fig. 6D).

“One-Size-Fits-Most” Montefiore Endovascular Grafting System

The Montefiore Endovascular Grafting System (MEGS; see Fig. 1) allows the graft to be customized during surgery, thus eliminating the need for preoperative measurement of vascular anatomy and graft fabrication. The MEGS has been previously described. ^34,39 In brief, it consists of a main aortounifemoral graft and a contralateral common iliac artery occluder. The main graft is constructed by suturing a Palmaz stent (P5O14 or P4014, Cordis, Warren, NJ) to a standard wall-thickness expanded polytetrafluoroethylene (ePTFE) graft (6 mm × 40 cm; IMPRA, Tempe, AZ). ^34,39 This stent-graft combination is mounted onto a large percutaneous transluminal angioplasty balloon (Maxi LD 25 mm × 4 cm; Cordis) and then inserted into a 16F sheath (Cook). The occluder device (for occluding the opposite common iliac artery) is constructed by attaching a Palmaz stent (P4014 or P308) to a 4-cm-long ePTFE graft that is closed at one end by ligatures (Fig. 1). This occluder is also mounted onto a percutaneous transluminal angioplasty balloon and then inserted into either a 12F or 16F sheath. These grafts are prefabricated and kept sterile for emergent use.

The MEGS was used under an FDA-approved investigational device exemption and also under institutional review board approval. Written informed consent was obtained from each patient or a family member.

Deployment Technique

Coil embolization of the hypogastric artery ipsilateral to the side of the MEGS main graft insertion was performed before or after its deployment (see Fig. 2). The MEGS delivery system was inserted into the aorta over a superstiff wire placed by means of the femoral sheath in the ascending thoracic aorta. If an infrarenal occlusion balloon had been deployed, the brachially placed balloon was inflated in the supraceliac aorta, and the infrarenal occlusion balloon was deflated and removed just before insertion of the MEGS main graft. Once the graft was inserted into the proximal aneurysm neck, the delivery sheath was retracted and an angiogram was repeated by means of a pigtail catheter introduced through the contralateral femoral artery to confirm appropriate positioning with regard to the renal arteries. Inflation of the large deployment balloon expanded the stent and secured the main graft in the proximal neck. By varying the inflation pressure in this balloon, the proximal stent was expanded to accommodate a wide range of proximal neck diameters, ranging from 20 to 28 mm (see Fig. 3). In each case, the length of the graft was 40 cm, so that the distal end of the graft always emerged from the introduction arteriotomy site (see Fig. 4A). The graft was then cut to the appropriate length and hand-sewn endoluminally in the common femoral or distal external iliac artery (see Fig. 4B). For AIAs, the main graft was placed in the proximal neck and the occluder device was placed in the opposite common iliac artery, thereby preserving at least one hypogastric artery. In addition, a femorofemoral bypass was performed (see Fig. 2). For isolated iliac artery aneurysms with a proximal common iliac neck, the main EVG was placed from that neck to the common femoral or external iliac artery, where it was secured by an endovascular anastomosis or a second stent, respectively (Figs. 7 and 8).

Figure 7. (A) Preoperative computed tomography (CT) scan of a 85-year-old man who came to the emergency room with severe hypotension, syncope, and excruciating abdominal pain. A CT scan revealed a large right common iliac aneurysm and a large retroperitoneal hematoma (H). (B) Postoperative CT scan shows a proximal stent (P) deployed within the proximal neck. The retroperitoneal hematoma (H) was subsequently evacuated through a limited retroperitoneal incision.

Figure 8. This patient came to the emergency room after developing abdominal pain and experiencing syncope. (A) Preoperative computed tomography (CT) scan showed a large ruptured iliac pseudoaneurysm of unknown cause. (B) Postoperative contrast-enhanced CT scan revealed exclusion of the large aneurysm. The intravenously injected contrast was confined within the endovascular graft (E). C, contrast remaining from intraoperative angiogram. This large hematoma was subsequently evacuated through a limited retroperitoneal incision.

Standard Open Repair

A standard transabdominal approach was used for patients who did not fit the EVG protocol. ⁴⁰ If the transbrachial supraceliac balloon was used, it was replaced with an infrarenal aortic clamp in a standard fashion. A tube graft was preferred. Axillobifemoral bypass was performed in one patient with a suspected mycotic aneurysm (see Table 2).

RESULTS

Seventeen ruptured AIAs were treated with the MEGS device (see Figs. 6 and 7), and two ruptured AIAs and one ruptured AAA were treated with a Corvita tube graft (Corvita Corp., Miami, FL) (see Fig. 8). Initially, the brachial wire was placed selectively; however, during the past 18 months, in all patients taken to the operating room for presumed ruptured AIAs, we routinely placed the brachial guidewire. The median time required to place the brachial guidewire in the descending aorta was 20 minutes (range 7–30 minutes). Because this maneuver was performed under local anesthesia, blood pressure was maintained at more than 50 mmHg in all patients during this step, and the time required for guidewire insertion did not result in any adverse outcome.

Five patients did not fulfill the inclusion criteria for an EVG repair, and they were treated with open repair. The reason for their exclusion was lack of a proximal neck (n = 1), bilateral long iliac artery occlusions (n = 1), abdominal compartment syndrome requiring laparotomy (n = 2), and mycotic aneurysm (n = 1). In all patients undergoing open repair, a brachial wire was introduced percutaneously under local anesthesia, but only two required supraceliac balloon inflation. These two patients had a hostile abdomen and the brachial proximal balloon occlusion rapidly stabilized their blood pressure. In both, the balloon was deflated and was replaced by an infrarenal standard aortic clamp as soon as the proximal neck was surgically dissected. Supraceliac balloon occlusion was required in seven patients treated with an EVG. In all nine cases, inflation of the proximal occlusion balloon resulted in an immediate rise in the blood pressure to more than 100 mmHg. Other surgical details are shown in Table 3.

Table 3. SURGICAL DATA

EVG, endovascular graft; MEGS, Montefiore Endovascular Graft System; PRBCs, packed red blood cells.

The EVG was successfully deployed at the target site in all patients, resulting in exclusion of the ruptured AIAs, cessation of bleeding, and absence of an endoleak (see Fig. 6E–J). This was accompanied in each patient by the resolution of preoperative symptoms.

Two patients treated with an EVG died within 30 days of the procedure (Table 4). One died in the operating room. This patient was diagnosed before surgery with an extensive acute myocardial infarction as well as a ruptured AIA. His death was due to left ventricular dysfunction leading to cardiac arrest. Another patient died 3 weeks after surgery of multiple organ failure. There were no deaths in the patients who had conventional surgery.

Table 4. DEATH AND COMPLICATIONS

EVG, endovascular graft.

Major complications included one case each of distal embolization requiring embolectomy, stroke, respiratory failure resulting in prolonged intubation, acute renal failure requiring temporary hemodialysis, and ischemic colitis. The latter resolved with conservative treatment.

Of the 20 patients treated with an EVG, evacuation of the abdominal hematoma was performed through a limited retroperitoneal incision in two patients and by a transperitoneal incision in one (see Fig. 6I). These large hematomas developed as a result of the initial rupture. This was performed for an overt abdominal compartment syndrome in one patient and prophylactically in two.

Of the 23 surviving patients, 7 died during a mean follow-up of 13 months. All deaths except one were unrelated to the surgery or ruptured AIA and were secondary to either cardiac disease (n = 4) or progression of preexisting malignant disease (n = 2). One patient died of bleeding from her recurrent aortoenteric fistula 18 months after surgery. A postmortem study revealed erosion of the Corvita graft secondary to infection.

Two EVGs failed during the follow-up period. One graft occluded 2 months after surgery and an axillofemoral bypass was required. However, this patient’s aneurysm continued to be excluded from the circulation until she died of her preexisting cervical cancer. Another patient developed a late endoleak 12 months after surgery. This endoleak resulted in the recurrence of severe abdominal pain. This patient was successfully treated with the insertion of a second MEGS device, and she continues to do well 30 months after the initial repair.

DISCUSSION

Our experience to date with the EVG treatment of ruptured AIAs demonstrates the feasibility of this approach even in unstable patients. Moreover, the insertion of a transbrachial guidewire and catheter into the supraceliac aorta under local anesthesia allows us to determine whether EVG repair is suitable and provides a means for proximal aortic control that can be used whether or not EVG treatment is possible. We believe the combination of this method of proximal control with the use of EVGs will improve the overall treatment outcomes for this highly morbid condition.

Our endovascular approach to ruptured AIAs has evolved as we gained experience, so some of the patients treated in the early phase were selected on the basis of relative hemodynamic stability. Therefore, our results cannot be simply compared with those previously published for ruptured AIAs. Despite this limitation and the small size of our series, we believe that our approach to the management of ruptured AIAs represents a better way to treat them and warrants further clinical trials. This belief is supported by the 8% death rate in our 25 patients. This compares favorably with the 24% to 70% rates for surgically treated ruptured AAAs 2–17,18–22 and AIAs. ^41,42

Endovascular grafts have been used to treat a variety of arterial lesions, including elective AAAs, thoracic aneurysms, iliac artery aneurysms, occlusive disease, and traumatic lesions, with promising short-term and midterm results. ^23–32 Although EVGs are theoretically appealing because of their less invasive nature, reports of their use in the treatment of ruptured AIAs have been limited to a few isolated cases and our initial series. 25,33,34,36

There are two reasons for this limited use of EVGs in patients with ruptured AIAs. First, most EVGs require thorough preoperative imaging and measurements of arterial anatomy to construct or procure the appropriately sized graft, in terms of both diameter and length. 24,30–33,36 This process usually takes several hours to weeks, which made the use of EVGs in the treatment of ruptured AIAs impossible or difficult. However, Yusuf et al ³³ have constructed a custom graft for the treatment of ruptured AIAs based on measurements obtained from the preoperative computed tomography scans. The entire process, including the time to sterilize the graft, required several hours; therefore, their patients, as well as those in other reports (including our own), ^25,36 included only those who could tolerate this delay. Because several hours of delay in performing a repair may carry a significant risk for these patients, we believe that EVG repair is usually justified only if it can be performed without such delays.

Unlike grafts that require thorough preoperative measurement, we have developed a graft that can be customized during surgery, largely eliminating the preoperative imaging (including a preoperative computed tomography scan) and measurements as well as the preparation of the graft for the specific patient. This unique graft, coupled with the use of a brachially placed wire and supraceliac aortic occlusion balloon, has minimized delay and enabled us to treat successfully even hemodynamically unstable patients with AIAs.

The second reason for the limited use of EVGs in this acute setting is that graft insertion may require a longer time to achieve proximal arterial control compared with standard surgical approaches, in which an infrarenal or supraceliac aortic clamp may be placed within 10 to 20 minutes after initiation of the procedure. In our earlier report, we described the selective use of a transbrachially placed supraceliac aortic occlusion balloon. ³⁴ Encouraged by this experience, we broadened the insertion of at least the guidewire into the descending thoracic aorta and currently we routinely place this before the induction of general anesthesia. This technique of obtaining balloon proximal aortic control is not new and was described as early as 1954. ^43,44 In 1977, Anastacio and Ochsner ⁴⁵ first described the use of fluoroscopic guidance and also the brachial artery approach for obtaining balloon control. We were the first to describe the use of and the potential benefits of this technique to facilitate EVG repair. ³⁴

Another unique aspect of our approach is the placement of only the guidewire on a routine basis. Only when it is needed is the balloon placed and inflated, and this can be done in a few minutes. The reason we do not place the balloon catheter routinely is because many patients do not require it. In addition, placement of the balloon catheter requires a large introducer sheath (8–14F). However, because a sudden drop in blood pressure below a lethal level after the induction of general anesthesia is unpredictable, and also because it takes much longer to deploy the EVG thorough an open femoral arteriotomy compared with standard supraceliac clamping, EVG repair without the use of this guidewire appears unwise. ^46,47 Moreover, the guidewire facilitates the arteriography that determines suitability for an EVG repair. Finally, we found that this maneuver is also useful even in patients who cannot undergo EVG repair. Surgical supraceliac clamping in experienced hands can be achieved within 10 to 20 minutes after the induction of general anesthesia. Although the placement of the brachial guidewire and the occlusion balloon may or may not take longer, there is an important difference between the quality of these two approaches. A surgical clamp can be applied only under general anesthesia and after opening the abdomen, both of which may exacerbate circulatory collapse. In contrast, endovascular balloon occlusion can be achieved without induction of general anesthesia and laparotomy. Thus, any extra minutes required to obtain endovascular control are probably justified, and endovascular control appears to be safer than a purely surgical approach, which can also cause many inadvertent venous and other organ injuries. 4,6–9,11 These iatrogenic injuries are due to the distortion of intraperitoneal and retroperitoneal anatomy by the hematoma and the urgent nature of the dissection. These problems do not complicate the insertion of the brachial guidewire or the balloon. We therefore believe that brachial guidewire insertion and proximal balloon control are important assets, even for patients who require an open repair.

There has been controversy over the value of intravenous fluid resuscitation in the management of patients with ruptured AAAs and AIAs. Correction of the intravascular volume deficit in patients with hemorrhagic hypotension is often recommended once intravenous access has been established. This premise is based mainly on several earlier animal studies showing that aggressive volume resuscitation with autologous blood was necessary to treat hemorrhagic shock. ^48–50 It is believed that intravenous administration of blood, crystalloid, or a combination of these agents sustains tissue perfusion and vital organ function. The concept of restoring normal blood pressure in hemorrhagic shock seems reasonable. However, the applicability of the controlled hemorrhage animal model to the clinical setting of a ruptured AIA is questionable.

Several recent studies indicate that aggressive volume resuscitation may be detrimental in the presence of an uncontrolled vascular injury. ^51,52 In all these studies, the authors showed that in the presence of an arterial injury, attempts to restore blood pressure by rapidly administering blood or crystalloid resulted in an increased volume of hemorrhage and a markedly higher death rate. In an earlier study, Milles et al ⁵³ concluded that in uncontrolled arterial hemorrhage, maintaining the blood volume tends to perpetuate the bleeding and results in a higher death rate than if no therapy were given immediately after the hemorrhage. It has been proposed that hypotension slows bleeding and allows local clot formation and tamponade of the bleeding. In contrast, the increased blood pressure caused by volume expansion may lead to dislodging of clot and increased bleeding. In addition, hemodilution from crystalloid administration may increase coagulopathies and third-space expansion. ⁷

The hypothesis that hypotension from restricted volume resuscitation may be beneficial was tested in a recent study ⁵⁴ that focused on the effect of restricted resuscitation in hypotensive patients with penetrating injuries. The study showed that delay of aggressive fluid resuscitation until surgical control of bleeding had been obtained improved the survival rate and shortened the hospital stay. Lawler ⁵⁵ first suggested that preoperative transfusion may be contraindicated in the treatment of ruptured AAA based on the assumption that increased blood pressure can be responsible for further exsanguinating hemorrhage. Crawford ⁷ in 1991 also advocated restricting volume resuscitation in ruptured AAAs, with maintenance of the blood pressure in the range of 50 to 70 mmHg.

In view of this evidence, we believe that intravenous fluid resuscitation should be limited to patients in whom the blood pressure fell to less than 50 mmHg. Resuscitation should be terminated as soon as the blood pressure exceeds 50 mmHg. In our experience, this limited fluid resuscitation has been effective, and all our patients could safely undergo brachial wire placement under local anesthesia.

One advantage of EVG repairs for ruptured AIAs is the potential to reduce blood loss. In our experience, EVG repair was accomplished with relatively small amounts of measured blood loss (median 400 mL, range 100–2,000 mL), although additional blood must have been sequestered in the retroperitoneum. In contrast, blood loss for open ruptured AAA repair has ranged from 2,600 to 8,000 mL. ^3,12,13 In elective AAA repairs with EVGs, the amount of blood loss has also been reported to be less than that accompanying open repair. ⁵⁶ However, the value of diminished blood loss may be far greater in patents with ruptured aneurysms, because these patients have already lost a significant amount of blood from the rupture, and further blood loss can contribute to disseminated intravascular coagulopathy, which can be a devastating complication, particularly with the extensive retroperitoneal dissection associated with an open repair. The reduced blood loss of EVG repair was due to the fact that it did not release the tamponade effect within the retroperitoneum and other surrounding structures. In addition, back-bleeding from branch arteries, bleeding from anastomotic suture lines, and bleeding from the iatrogenic venous injuries, all of which are main sources of blood loss during standard repair, can be eliminated during EVG repair.

Finally, EVG repair is performed without laparotomy, which may be partly responsible for the hypothermia often encountered during standard open repair. ⁵⁷ Hypothermia can exacerbate poor cardiac function and coagulopathy and contribute to death after surgical repair. EVG use can minimize hypothermia by avoiding laparotomy, and this may be an additional advantage.

From our experience, we conclude that the approach of placing a transbrachial guidewire and catheter into the supraceliac aorta, obtaining an arteriogram, and using an EVG to exclude the rupture site whenever possible is a feasible way to treat many if not most patients with ruptured AIAs. This approach facilitates the use of a supraceliac balloon catheter to obtain proximal control if needed, but such control is unnecessary in many patients with a ruptured AIA if fluid resuscitation is restricted and the blood pressure is allowed to remain low. Balloon proximal control may be essential in some patients suitable for EVG treatment and may be a valuable adjunct in those who require open repair. Use of EVG repair and these other image-guided, catheter-based techniques may improve treatment outcomes in patients with ruptured AIAs.

Discussion

Dr. James S. Yao (Chicago, Illinois): This is a somewhat selective series, but even so, a 6% mortality rate for treatment of ruptured aortic aneurysm is very impressive. I have three questions.

One is the use of CT scan. The CT can detect the thrombus where an angiogram is not able to do so. I want to know whether you have any concern about not using CT scan in your series.

The second question is about endovascular graft. I know that this is a somewhat homemade “one size fits all” prosthesis. I would like to know, if I want to do the procedure, do I need FDA approval? Also, do you have any plan to commercialize the graft?

Finally, since this is a very labor-intensive technique, how do you foresee it being used in a community hospital at 3:00 am?

Presenter Dr. Frank J. Veith (New York, New York): You are correct about the CT scan being absolutely essential to rule out clot in the aneurysm neck. However, the percentage of cases that have clot in the neck in our experience with elective aneurysms is fairly small. It certainly has not been a problem in this series of ruptured aneurysms. It is a concern, but in the urgent setting we think it is a risk worth taking.

As far as the surgeon-made graft is concerned, we have performed this protocol—indeed all of our use of the MEGS graft—under an investigator-sponsored FDA IDE. If one wants to make their own graft and use it, it is wise to have such an IDE. It is a little easier for us to obtain an IDE than for industry to do so; however, it is still a big job. If one does not have an IDE currently, then one is taking a considerable risk.

Your third question related to the community hospitals. If one does not have the expertise necessary to insert a brachial balloon to help with a standard open surgical repair, I think it is well worth the effort to acquire these skills and the equipment to insert such a balloon. It can be very useful. Obviously, we think all these techniques will be adopted more and more in endovascular centers, and will ultimately find their way to the community hospital.

We are making an effort to commercialize our graft, but that is probably a year or two away.

Dr. Anthony J. Comerota (Philadelphia, Pennsylvania): As you mentioned, epidemiologic studies have shown that the mortality for both elective and emergency operations for abdominal aortic aneurysms are unchanged over the past several decades, and that will be confirmed in a current epidemiologic study at Eastern Vascular later this year. If this technique and outcome can be reproduced by other centers it represents an enormous contribution in these very high-risk patients. I have four short questions, Dr. Veith.

How many of your patients were truly hypotensive, which is typical of the patients we see with unstable ruptured abdominal aortic aneurysms? That may play on your results.

Ischemic colitis is more common after ruptured aneurysm repair than elective repair. And in those patients with large hematomas, you have chosen to intentionally exclude at least one internal iliac and, of course, eliminate the possibility of reimplanting the inferior mesenteric artery. Did you see ischemic colitis? And how would you manage it?

You chose an aortofemoral bypass because of the nature of your graft. Do you anticipate that you will convert to an aortoiliac in the future?

Finally, could you address whether you had any difficulties passing a relatively large brachial balloon catheter in a potentially constricted brachial artery in a patient in shock?

Dr. Veith: With regard to the hypotensive state, in the beginning we started out doing only patients in whom we could get a CAT scan and plan the cases. These patients had contained ruptures. The last group of patients that Dr. Ohki presented were all comers to our emergency room who survived long enough to be taken to the OR, meaning 10 or 15 minutes in transit. Moreover, the patient that Dr. Ohki showed and several others would probably not have survived an open operation in anybody’s hands. Such cases were truly quite dramatic. Obviously, there were also some less dramatic cases, but all of these patients were true ruptured aneurysms, and within the group there were two or three patients who had free rupture into the peritoneum.

With regard to the ischemic colitis, it is clearly a big concern. In this series of patients, we had only one patient who had a low-grade ischemic colitis that resolved with conservative therapy. As you probably know, we frequently occlude the hypogastric artery on one side and occasionally on both sides. Surprisingly, the number of bowel problems, i.e., ischemic colitis problems, with these hypogastric occlusion patients have been quite limited and innocuous. We currently believe that unilateral and even bilateral hypogastric occlusion is an acceptable risk to take in certain patients with large bilateral common iliac aneurysms.

We do not plan to go to a bifurcated system because we think “one size fits most” is the way to go. Whether or not you could substitute an aortouniiliac graft is a moot point. We like our system. One does not have to worry about length or inserting a distal stent to fix the graft. Therefore, we will stick with the aortofemoral graft. It has advantages.

The large brachial sheath, the 14-French sheath, is a concern. All these patients who had this sheath required an open repair of the brachial artery. With this, however, we have had no problems with the big sheath and the balloon. Of course, we do have concerns about such a large sheath in the brachial artery, and hope that better balloons will permit use of smaller sheaths.

Dr. Thomas F. O’Donnell, Jr. (Boston, Massachusetts): For ruptured aneurysms, this approach is truly unique and on the surface logical, because the sudden decompression of intraabdominal pressure associated with opening the abdomen is avoided with the endovascular approach. The accompanying drop in blood pressure may not occur. In addition, the subsequent third-space requirement due to the transabdominal approach may be minimized. Could you describe for us the postoperative fluid requirements in the endovascular cases? Since these cases were discharged in 6.2 days, I would think that volume of fluid infused might be less than in the typical open case.

I wanted to follow up on the earlier discussants’ comments on mortality results. Critical to evaluating mortality results in any ruptured aneurysms series is the phenomenon of pretreatment selection bias where only a proportion of patients are offered surgery. This has been debated recently in the vascular literature around the Seattle series. What portion of the ruptured aneurysm series does this represent, and is it comparable for the open and endovascular groups?

Finally, your Phase 1 study of the endovascular approach shows that it may be superior to the open approach. Through your previous association and teaching role in endovascular surgery, you have a network of surgeons who are skilled in these capabilities. Are you planning to do a randomized control trial to prove or disprove efficacy of the approach, as you did with PTFE versus vein for infrainguinal bypass?

Dr. Veith: The fluid requirement has varied. Obviously these patients have a considerably larger requirement for blood compared to the elective endovascular aneurysm repairs because they had a lot of blood in their retroperitoneum. They also have not been as sick postoperatively as the standard open repair. But some of them have been fairly sick, as reflected by the fact that they require 6 days in the hospital, whereas the standard endovascular AAA repair requires 1, 2, or 3 days. In general, the fluid requirements are less, but they are not negligible.

The selection bias in our series is real. In the first batch of cases, which I think numbered 12, we clearly selected the patients on the basis of their being stable enough to be worked up. And those patients included a number of contained ruptures. But as Dr. Ohki mentioned, the last 15 cases have been unselected. In there, the first maneuver was to take them to the operating room and to use the protocol we have described. Whether or not others will be able to reproduce our results, we will have to see—we are optimistic, but it remains to be seen.

We do not plan to do a randomized prospective study, although I suppose somebody may in the future do that. Our next phase will be to do a multicenter study, if we can get to the point where we can have our graft commercially available. And that would be the Phase 2 study we spoke about. It would be a multicenter study using our graft for all ruptured aneurysms.

Dr. Lazar J. Greenfield (Ann Arbor, Michigan): Many of my questions have been answered already, but I would like to pursue the issue of the brachial occlusion catheter a little bit more, by asking how long you would be willing to pursue this in a hypotensive patient if you had some difficulty in positioning the catheter, since I believe that is one reason why it has not been more widely adopted.

The other area has to do with your system, which is unique in its ability to adapt to any size of a proximal aorta. What are the limits of this application? In particular, is it possible to do this with a tapered rather than a straight neck?

I think that this is a real contribution in the fact that you were able to apply it so broadly, finding 20 of 25 cases suitable. But how does this high level compare with your elective aneurysm candidates? Do you find the same ability to insert in them? And do you use the same system?

Your outcomes were impressive, with one death from an MI and one from multiple organ failure. Were there any technical problems with prolonged underperfusion that contributed to the multiple organ failure?

The technical consideration of the large catheter in the brachial artery does raise the question in a patient who is not anticoagulated of whether you might have to do a thrombectomy at that level. Was that necessary?

You did have one occlusion in follow-up. Since that patient subsequently died of cancer, were you able to examine the graft at autopsy?

The development of a late endoleak in your patient after 12 months is one of the nagging problems with the technique. Since this can occur from a patent IMA or a lumbar, are you considering any further occlusive approaches to these vessels at the time of graft placement?

Finally, we are all interested in the fate of the excluded aneurysm. You have been able to measure the pressures in the occluded segment after proximal graft deployment. How do you use this information? What has happened to aneurysm size at late follow-up?

Dr. Veith: Dr. Greenfield, you asked lots of questions, and they are all good ones.

Obviously, rather than inserting a brachial balloon in a patient who was crashing at the time he was in the OR, one might be tempted to use the more standard approach. However, in our patients, when we used the “hypotensive hemostasis” (which is not a new concept with us but has been advocated by Stan Crawford, Gerry Shaftan, and Ken Mattox in other circumstances), i.e., when we left the patients largely unresuscitated and let their blood pressure fall to 50, we have almost always had the opportunity to take the 10 or 15 minutes to place the brachial wire. Then the patient is put to sleep. And then they usually crash, or often they crash. At that point with the wire in the aorta, one can fairly quickly place a larger sheath and the brachially placed occlusion balloon in the thoracic or supraceliac aorta. We have not yet had a problem where we have had to abandon this approach, although I guess it could occur. Occasionally, we have had to make a cutdown in the brachial artery, but that was only one time when we had trouble with the percutaneous brachial puncture.

The limitations on the size of the neck of the aneurysm are real. Our graft only goes to a maximum of 28 mm. So if the neck is bigger than that, obviously our graft is not suitable. Whether or not other grafts will be developed that are larger remains to be seen.

The tapered neck is not as much of a problem as the flared neck. And if we saw a flared neck or a very short neck on the arteriogram we would probably use the brachial wire and a balloon as necessary, and then do an open repair. Two thirds of the patients in our unselected series could have an endovascular graft repair and one third required an open repair, either because of neck anatomy or because of an abdominal compartment syndrome.

Multiple organ failure—the patients who had it were desperately ill. They usually had a complicated repair. Obviously the complexity of the repair and its duration contributed to their multiorgan failure. But they were pretty sick candidates to begin with.

We did not get an autopsy in the one case that died, unfortunately. He succumbed in another hospital.

We do not worry very much about the branch lumbar or IMA (type 2) endoleaks. The patient that developed a late endoleak had a type 1 leak from the proximal anastomosis and Dr. Ohki was able to insert a second graft successfully. That patient continues to do well over a year later.

The business about pressure in the sac. We do measure it intraoperatively. If the pressure does not fall, the aneurysm is not excluded. In our elective experience, the pressure in the sac after graft placement often falls to the 40- or 50-mmHg level or less. This means that the straight-line arterial flow is excluded from the aneurysm. If the pressure is high and pulsatile, there is a major leak.

Some of these aneurysms have gotten smaller following the procedure as time goes on. But with our graft, for some reason, the aneurysms do not tend to shrink as much as do those that had a bifurcated fully stented graft.

Dr. Gregorio A. Sicard (St. Louis, Missouri): This definitely will become a landmark paper in the treatment of ruptured aneurysms. My question is the following: What do you do with your ipsilateral hypogastric artery? Normally when we do this in elective situations we embolize them. Obviously this is not the type of patient that you want to embolize. Secondly, why not go through the femoral and avoid the sheath issues in the brachial? And then you still can put in a second catheter for an arteriogram.

Dr. Veith: The ipsilateral hypogastric—if there is any component of iliac dilatation, we make an effort to embolize the hypogastric at some point in the case. Once, we were able to pass the catheter and wire outside the graft and embolize it even with the graft in position. However, that may not be an easy thing to do. If the common iliac is small, in a number of the cases we have just allowed the graft to cover the orifice of the hypogastric and have not felt compelled to embolize it.

Your second question, the femoral approach for the balloon. First of all, there is a tendency—and we have all used balloons in the past—the tendency is, as the patient gets resuscitated and his pressure comes up, the balloon is displaced distally with the force of the aortic flow. With the brachially introduced balloon, you have the catheter to hold it in place. Actually, I also think it is easier in patients with tortuous iliacs to insert the balloon at the beginning from above. That is just an opinion. I believe others like Brian Hopkinson have inserted a balloon through a femoral approach. Of course, the ideal thing is not to use the balloon at all. In many patients, it has been possible with hypotensive hemostasis to insert the graft without inflating the balloon.