Hybrid Treatment of a Ruptured Diverticulum of Kommerell

CASE REPORT

An otherwise healthy 82-year-old woman with no medical history presented with sudden onset of thoracic pain radiating to the right arm, neck, and upper back region. The patient was fully conscious at presentation. Blood pressure and heart rate were 138/82 mmHg and 115 beats per minute, respectively. An electrocardiogram was normal. Suspecting an aortic dissection, computed tomographic angiography (CTA) was performed, which showed a large hematoma with contrast in the upper mediastinum contained by the parietal pleurae. With the tentative diagnosis of a DeBakey type I aortic dissection, the patient was transferred to our hospital. Closer examination of the CTA (Figure, A and B), however, revealed that the actual diagnosis was not an aortic dissection but a rupture of the aneurysmal origin of a right aberrant subclavian artery: a ruptured diverticulum of Kommerell (DoK).

OPEN IN VIEWER

Figure

(A) CTA at presentation. 1: right carotid artery, 2: left carotid artery, 3: left subclavian artery, E: esophagus, T: trachea, DA: descending aorta, ARSA: aberrant right subclavian artery, R: rupture. (B) Three-dimensional CTA reconstruction showing the left-sided aortic arch, the aberrant right subclavian artery, and the mediastinal hematoma due to rupture of the diverticulum of Kommerell. AA: ascending aorta, DA: descending aorta, LSA: left subclavian artery, ARSA: aberrant right subclavian artery, LVA: left vertebral artery, RVA: right vertebral artery, RCCA: right common carotid artery, LCCA: left common carotid artery, H: hematoma. (C) Three-dimensional CTA reconstruction showing the stent-graft in the descending thoracic aorta, covering both subclavian orifices; the right-sided carotid-subclavian bypass; and the vascular plug located in the proximal ARSA. 1: right carotid-subclavian bypass, 2: vascular plug, 3: distal ARSA, 4: stent graft, RVA: right vertebral artery, LVA: left vertebral artery.

Because of imminent circulatory instability, the patient underwent immediate combined open/endovascular surgery in a standard operating room equipped with a mobile C-arm and image intensifier (Philips Medical Systems, Best, The Netherlands). During the procedure, the patient had permissive hypotension and was not heparinized; a diagnostic angiography catheter was introduced via the right common femoral artery for intraoperative imaging.

We decided to cover the rupture site with an aortic stent-graft. To obtain sufficient proximal sealing, both subclavian artery origins had to be covered, thus creating retrograde flow in the basilar and both vertebral arteries. To preserve antegrade flow in the basilar artery, a right-sided 8mm Dacron carotid-subclavian bypass procedure was done through 2 separate incisions, as it was anticipated that the right subclavian artery would be inaccessible via a supraclavicular approach due to its abnormal course. The DoK rupture was then covered with a 40-mm-diameter, 20-cm-long thoracic endograft (Valiant Captivia; Medtronic CardioVascular, Santa Rosa, CA, USA) via the left common femoral artery. A 16-mm vascular plug (Amplatzer; AGA Medical, Golden Valley, MN, USA) was placed in the right subclavian artery, proximal to the origin of the right vertebral artery, thus eliminating retrograde blood flow through the right subclavian artery to the rupture site. The plug was positioned as close to the aneurysm as possible to minimize the risk of future type II endoleak from subclavian side branches. Access to the right subclavian artery was established just proximal from the carotid-subclavian bypass anastomosis. Completion angiography showed a successful exclusion of the ruptured aortic segment and antegrade flow in the right vertebral artery, without signs of endoleak.

Two days after surgery, a decrease in hemoglobin and circulatory instability were noted. CTA (Figure, C) nevertheless did not reveal any sign of active bleeding or endoleak. A pleural effusion in the left hemithorax was causing respiratory distress, so a pleural drain was placed; 600 mL of serosanguineous fluid was collected. A routine CTA at 3 weeks postoperatively showed a type II endoleak from an aberrant right bronchial artery. However, the volume of the mediastinal hematoma had markedly decreased. Recurrent pleural serosanguineous fluid buildup was again aspirated. It was hypothesized that these recurrent pleural effusions were most likely caused by lysis of the mediastinal hematoma and temporary blood loss from the type II endoleak. Considering the stable hemodynamic condition of the patient at this time, a wait and see policy was maintained. The patient was discharged 3 weeks after surgery.

Unfortunately, 6 weeks after the operation, the patient's general condition deteriorated, requiring hospital admission. The patient had an episode of melena and developed anemia. CTA did not show any complications related to the procedure. However, esophagoscopy revealed 3 annular ulcers, suggestive of ischemia with necrosis and perforation. Because reasonable therapeutic options were lacking, it was decided to start palliative treatment, after which the patient was discharged from the hospital. The patient died of sepsis 8 weeks after surgery.

DISCUSSION

The diverticulum of Kommerell is a rare anomaly of the descending thoracic aorta, defined as an aneurysmal enlargement of the proximal portion of an aberrant right subclavian artery (ARSA). An ARSA is present in ~0.4% to 2% of all individuals, ² and ~60% of these patients have a DoK. ³ Dr. Kommerell described this anomaly in 1936 when a patient suffering from dysphagia was found to have a left aortic arch and an aberrant right subclavian artery. ⁴

Most patients with DoK remain asymptomatic, as the trachea and esophagus are usually not constricted by vascular structures. However, when the right and left carotid arteries arise from the aortic arch together or in close proximity to each other, symptoms may arise due to external compression of the esophagus and trachea from the ARSA dorsally and the carotid bifurcation ventrally, as was the case in our patient. ⁵ However, our patient never experienced any complaints of dysphagia. During adulthood, ~5% of patients with an ARSA have symptoms due to the development of atherosclerotic rigidity and tortuosity, especially when the ARSA originates from a diverticulum. DoK is seen more often in subjects with a right aortic arch and an ARSA than in those with a left aortic arch and a left aberrant subclavian artery.^6,7

Progressive aneurysmal and atherosclerotic degeneration frequently occurs. There is a marked propensity toward acute rupture, which is fatal in up to half of the cases. ⁸ Elective open surgical repair requires thoracotomy to gain control of the retroesophageal part of the ARSA; the operative mortality rate is ~25%. ¹⁰ Two cases have been described where a DoK was electively excluded with a stent-graft in combination with bilateral carotid-subclavian bypasses,^9,10 but to our knowledge, no report has been published of a combined open and endovascular procedure to treat a ruptured DoK.

Endovascular treatment of ruptured TAA is clearly less invasive and leads to a decrease in perioperative morbidity and mortality. ¹¹ Similarly, a reduction in operative risk would be expected from avoiding thoracotomy in favor of an endovascular approach when treating a ruptured DoK.

Any abnormal anatomy in these patients poses a challenge and may preclude simple stent coverage of the ruptured site. In their 2009 review, Ritvi et al. ¹² found very limited evidence suggesting that single subclavian artery coverage increases the risk of vertebrobasilar ischemia, spinal cord ischemia, and anterior circulation stroke. However, Holt et al. ¹³ found that left subclavian artery revascularization was associated with significantly better outcomes for the combined measure of death, stroke, and paraplegia after planned endovascular TAA repair. The Society for Vascular Surgery recommends revascularization of the vertebrobasilar vessels when possible. ¹⁴

In our patient, both the left and the right subclavian artery exited at the same level, necessitating exclusion of both origins in order to effectively cover the ruptured site. It is to be expected that coverage of both subclavian orifices will increase the risk of cerebral ischemic events, especially in the case of an incomplete circle of Willis. Considering the relative urgency of the situation at hand, we decided to perform a unilateral instead of a bilateral carotid-subclavian bypass to preserve antegrade flow in the basilar artery in order to expedite deployment of the aortic stent-graft and gain control over the ruptured site, as well as to reduce operation time.

An important postoperative problem in this patient was the type II endoleak from a bronchial artery in combination with the large mediastinal hematoma from the rupture. The endovascular treatment of such an endoleak is problematic, since there is no transluminal route to gain access to this vessel other than to cannulate it directly. Surrounding venous structures, pulmonary arteries, the left lung, and the small caliber of the bronchial artery make this virtually impossible. Alternatively, the endoleak can be approached via thoracotomy. Considering the technical difficulties (endovascular) and the considerable operative risk (thoracotomy), respectively, these options should probably be reserved for the hemodynamically unstable patient with angiographically confirmed hemorrhage from such an endoleak; in hemodynamically stable patients, a wait and see policy may be justified.

Large mediastinal hematomas are known to introduce a risk of serious and life-threatening complications in patients treated for ruptured TAA. Delayed ischemic esophageal perforation complicating ruptured TAA repair is rare and usually will lead to mediastinitis, graft infection, sepsis, and subsequent death. Literature on this topic is scarce. It is hypothesized that ischemia and necrosis are the result of compression by the mediastinal hematoma. Clinical features of this type of perforation are not well known, and this complication is usually not anticipated, leading to delayed diagnosis. It has been found after both open and endovascular thoracic aortic repair, and there are no measures that we know of to prevent this complication. ¹⁵ Therapeutic options are limited to expedient esophagectomy and omental transposition to cover the exposed graft. It seems obvious that the anticipated success of such extensive, high-risk salvage procedures is severely limited by the patient's condition, comorbidities, and age and should probably be performed only in selected patients.