Endoluminal Abdominal Aortic Aneurysm Repair Complicated by Intracardiac Guidewire Placement and Massive Transfusion

Endoluminal stenting, a recently introduced, less invasive approach to abdominal aortic aneurysm (AAA) repair, is an attractive alternative to open repair in high-risk patients with severe coexisting medical disease (1). We report a case of endoluminal AAA repair complicated by accidental intracardiac guidewire placement and massive hemorrhage, which illustrates the potential for the occurrence of major procedure-specific complications. It also confirms the value of intraoperative transesophageal echocardiography (TEE) in the rapid evaluation of acute hypotension in a critically ill patient. Case Report A 68-yr-old man with a mildly symptomatic enlarging 8.3-cm infrarenal AAA presented for elective endoluminal repair. Two previous attempts at open repair in another institution had been abandoned because of hypotension, myocardial ischemia, and pulmonary edema after the induction with standard combined general anesthesia and thoracic epidural techniques. On both occasions, laparotomy had commenced, and on the second occasion, the bowel was inadvertently perforated. A diagnosis of hypertrophic obstructive cardiomyopathy (HOCM) was subsequently made with postoperative echocardiography. Spiral computed tomography and angiography demonstrated the feasibility of endoluminal repair, although difficulties caused by narrowed iliac artery access and mild angulation of the aortic aneurysm neck were recognized. Because the rate of rupture of a large AAA within 5 yr approaches 40%, and the mortality for attempted repair of a ruptured AAA is high (2), elective endoluminal repair of the aneurysm was recommended after optimal management of HOCM was attained. At preoperative evaluation, no symptoms of angina, syncope, or arrhythmias were reported. Other history included mild chronic obstructive pulmonary disease and mild renal impairment with creatinine of 128 μmol/L. Medication consisted of sotalol 80 mg twice per day. Transthoracic echocardiography revealed asymmetric basal left ventricular (LV) hypertrophy and a mean pressure gradient across the LV outflow tract (LVOT) of 55 mm Hg which increased to 100 mm Hg with the Valsalva maneuver. Electrocardiogram (ECG) showed sinus rhythm, left axis deviation, and abnormal q waves and T inversion laterally. The procedure was conducted in the operating room. After insertion of radial and pulmonary artery catheters, anesthesia was induced with fentanyl and thiopentone and maintained with enflurane in oxygen and vecuronium. Preinduction cardiac index was 2.3 L · min−1 · m−2 and pulmonary capillary wedge pressure (PCWP) was 24 mm Hg at end-expiration. TEE monitoring (5-Mhz Omniplane TEE probe; Hewlett Packard Sonos 1000; Andover, MA) for the extent of obstruction was performed semiqualitatively from midesophageal 5-chamber views with the aid of color flow Doppler, as transgastric apical 5-chamber views and pressure gradients across the LVOT could not be obtained. To minimize LVOT obstruction during the procedure and maintain normal blood pressure, the goals of hemodynamic management were to maintain LV cavity size, as an index of preload, with fluids; decrease inotropy with enflurane; control heart rate with an esmolol infusion; and, prevent a decrease in afterload with a phenylephrine infusion. To decrease the risk of pulmonary edema, marked elevations of PCWP were avoided. Several unexpected technical complications were experienced during the course of surgery, leading to massive blood loss and prolonged operating time. An 18-cm long, self-expanding Talent bifurcated modular-stent graft with a 32-mm proximal end diameter was selected. Because the delivery system could not be advanced over a femoral wire through a narrowed right external iliac artery, an extraperitoneal approach via the right iliac fossa was performed to construct a Dacron graft anastomosis to the common iliac artery. The delivery system could not be advanced through the graft beyond midaortic level, however, and had to be removed, causing disruption of the anastomosis, which had to be revised, and a large blood loss. A guidewire was then inserted via a left brachial arteriotomy, passed through the descending aorta and aneurysm under fluoroscopic control, and retrieved from the distal graft. An attempt was then made to pass the device over the guidewire held with tension on each end to straighten the angulated aortic aneurysm neck (“body floss maneuver”). Progress of the device was again impeded, and re-examination of the wire showed fracture of the coating and bunching-up with graft passage. To exchange for a stiffer wire, a 140-cm long Hinck catheter was then passed from the left arm and through the thorax under fluoroscopic control to midaortic level. Before the catheter could be retrieved from the groin, acute marked hypotension, unresponsive to increased phenylephrine, was noted. An electrocardiogram showed nodal rhythm, and the PCWP was 28 mm Hg and unchanged. TEE examination at this point demonstrated looping of the catheter through the aortic valve into the LV and acute severe aortic regurgitation (Figure 1). The image intensifier, which had been focussed over the end of the catheter in the lower abdomen, was swung over the thorax to show that the midportion of the catheter had buckled down into the ascending aorta (Figure 2) after meeting distal obstruction in the abdomen. Sinus rhythm, cardiac output, and valvular function were quickly restored after removal of the catheter.Figure 1: Transesophageal echocardiography midesophageal ascending aorta long-axis view showing guidewire looped through open aortic valve into the left ventricle. LA = left atrium, ASC AO = ascending aorta, LVOT = left ventricular outflow tract.Figure 2: Chest radiograph showing intracardiac looping of guidewire inserted via a left brachial arteriotomy. PA = pulmonary artery.Eventually, the endoluminal graft was deployed over a superstiff Amplatz wire inserted from the groin via the Dacron graft. Completion arteriography showed successful deployment of the graft immediately below the renal arteries, with a small distal endoleak at the lower end of the right iliofemoral graft. After 7 h of operating time and blood loss of approximately 6.5 L, the patient was admitted to the intensive care unit. Intermittent left bundle-branch block was noted for 3 days but without increase in cardiac enzymes. He remained intubated until Day 7 because of mild fluid overload, atelectasis, probable over-sedation (to prevent stress and exacerbation of HOCM physiology), and confusion, and he was discharged home on Day 10. Because of persistent endoleak on the computed tomography (CT) scan, an additional iliac endoluminal graft was placed 2 mo later. Eighteen months later, the patient remained well, with computed tomography scan evidence of aneurysm control by the endoluminal grafts. Discussion The apparent minimally invasive nature of endoluminal repair has lead to the successful performance of the procedure under local anesthesia in the angiography suite (3). However, despite experience and improvements in technology, adverse systemic and local vascular events continue to be reported in up to 45% of patients undergoing endoluminal AAA repair in established referral centers (4). Procedure-specific complications include renal insufficiency as a result of vessel obstruction and contrast load, mesenteric artery obstruction, and distal embolization caused by manipulation of the device within the aneurysm. Although blood loss from leaky valves and the delivery systems is usually less than during open repair (5), trauma from femoral dissection and passage of the large delivery device may lead to vessel perforation and massive hemorrhage. Furthermore, as seen in this report, the procedure may not be entirely percutaneous, and recognition of the extent of blood loss may be impaired by the darkened room. Incidence of complications increases from 37% to 71% in patients with difficult complex aortic morphology (6). In addition, urgent conversion to an open procedure was associated with a mortality rate exceeding 40% in one series (7). Because of the lack of postoperative pain, risk of long-acting sympathectomy during an episode of major blood loss, and variable duration of surgery, most endoluminal procedures in our institution are performed under general, rather than regional, anesthesia, with direct arterial and central venous pressure monitoring and large bore IV access. In the case presented, the high-risk nature of the procedure in this patient with severe coexisting medical disease was anticipated, and the patient was transferred from the angiography suite to the operating room for more extensive monitoring. PCWP was elevated at baseline because of diastolic dysfunction and mitral regurgitation secondary to HOCM and proved to be a less helpful monitor of preload than LV cavity size during the massive blood loss. Because the narrow field of view of the image intensifier was focussed over the lower abdomen, the altered course of the guidewire in the chest during advancement was not noticed. Fortuitously, TEE, inserted primarily to monitor the extent of subaortic obstruction and LV cavity size, detected the intracardiac guidewire placement before the onset of irreversible hemodynamic deterioration. Confusion with other causes of hypotension was thus avoided. In conclusion, endoluminal AAA repair is a promising alternative technique to open repair with generally favorable short-term outcomes (5,8–10). While patient selection, device design, and deployment techniques for this procedure continue to evolve, it is important to be aware of the implications of procedure-specific maneuvers not seen with traditional open surgical techniques. The anesthesiologist must remain vigilant during planning and performance of the procedure and be prepared for rapid, large-volume blood loss and the resulting need for conversion to open repair. Patients with difficult aortic anatomy, in particular, need to be identified preoperatively to allow for more intensive monitoring in an appropriate setting. The authors thank T. Devine, FRACS, and K. Myers, FRACS, for helpful comments.