Aortic Dissection as a Complication of Celiac Plexus Block

A 52-yr-old man was evaluated for intractable abdominal pain secondary to unresectable gastric carcinoma involving the greater and lesser curvatures of the stomach. Two months before our evaluation, a gastroscopy revealed a bulky carcinoma. A computed tomography (CT) scan demonstrated the tumor to be locally invasive. His history was significant for allergy to contrast media, which produced an anaphylactic response.

The patient underwent a celiac plexus block while in the prone position. A 15-cm, 20-G block needle was directed from the patient's left side and positioned, with the aid of fluoroscopy, until the tip was anterior to the first lumbar vertebra. Transaortic needle placement was performed as described by Ischia et al. [3]with verification of passage through the anterior aortic wall as described by Feldstein et al. [4]Complete relief of his pain was attained after injecting 40 ml 0.5% lidocaine, using intermittent aspiration. Injection of an equal volume of absolute alcohol was planned with the same intermittent aspiration. After approximately 30 ml of alcohol was injected, blood was aspirated. The needle was advanced to reposition it anterior to the aorta. When pressure was exerted on the plunger of the syringe, the patient suddenly complained of generalized body pain, and aspiration again produced blood return. The procedure was aborted, and the patient was turned supine and treated with morphine for his pain, which gradually subsided over about 15 min. There was no evidence of peripheral vascular insufficiency; the patient voided about 45 min later, and a bowel movement occurred about 2.5 h later. Vital signs remained normal, and his abdomen was soft, nontender, and pain-free until 4 h after the injection, when he complained of abdominal pain, which became severe and excruciating over the next 24 h. His temperature increased to 38.5 degrees Celsius, and his abdomen became distended.

A noncontrast CT scan revealed gastric tumor extending throughout the central abdomen, emphysematous cholecystitis, and pneumatosis intestinales and/or gangrenous bowel. An emergency exploratory laparotomy revealed emphysematous cholecystitis, infarct of the entire duodenum and 0.91 m of proximal jejunum, purulent ascites, and thrombosed mesenteric veins. Mesenteric and hepatic arterial pulsations were described as present and good. Surgical resection was not possible. Postoperatively, mechanical ventilation was maintained, along with a morphine infusion of 50 mg/h, until his death on the 4th postoperative day.

A postmortem examination confirmed the presence of an infiltrating adenocarcinoma of the stomach and multifocal intraabdominal ischemic injury. Examination of the anterior aortic intima between the origins of the superior mesenteric and celiac arteries revealed a puncture mark consistent with needle penetration. This lesion was continuous with intramural dissection of blood in a small portion of aorta and the proximal superior mesenteric and hepatic arteries. The hepatic artery dissection was associated with thrombosis (Figure 1and Figure 2). The presence of dissection and vascular thrombosis in these areas resulted in bowel and liver infarction.

The incidence of major complications from neurolytic celiac plexus block was determined by Davies. [5]In 2,730 neurolytic blocks performed from 1986 to 1990, the overall incidence of major complications (paraplegia, bladder and bowel dysfunction) was 1 per 683 procedures. The report did not indicate the block techniques.

Transcrural [6,7]and transaortic [3]techniques were developed to minimize the neurologic complications occasionally reported with the retrocrural approach. Singler [6]and Hilgier and Rykowski [7]reported using the transcrural approach in 41 patients, with transient diarrhea as a side effect in some patients. In 148 patients reported to have received the transaortic approach, [3,4,8]transient orthostatic hypotension and diarrhea were reported as the main side effects in some. If the incidence reported by Davies [5]is reproducible, the number of patients from these articles may be too low to say these techniques are free from major side effects. For example, Matson et al. [9]reported a potentially fatal dehydration from diarrhea, and Traycoff et al. [10]described paraplegia after a transaortic neurolytic block.

Only one other case of aortic injury from celiac plexus block has been reported. [11]The block technique was not described. The procedure was performed for pain from chronic pancreatitis. In this case, a traumatic pseudoaneurysm resulting from the procedure was successfully repaired surgically using a Dacron graft through a thoracoabdominal approach.

The event leading to the dissection in our case was movement of the needle tip from its original position anterior to the aorta to within the aorta and the subsequent attempt to reposition it. Either may have created the tear noted at the autopsy, permitting some injection of alcohol into the lumen. The alcohol may have contributed to local tissue damage/necrosis, and blood, under pressure in the aorta, completed the dissection. In addition, the total body pain experienced by the patient may have been secondary to some intraarterial injection of alcohol. As increments of no more than 1 ml of alcohol were injected before each aspiration, intraluminal alcohol injection leading to mesenteric and hepatic artery thrombosis was considered but seems unlikely in view of the autopsy findings. The mesenteric and hepatic pulses noted at surgery were likely transmitted sensations from the aorta.

The reason for movement of the needle tip from anterior to the aorta to within its lumen is uncertain, as the needle was secured in position during the injection. Forty milliliters of local anesthetic was required to obtain relief, and anatomic distortion from the tumor as well as this volume of injectate may have shifted the needle tip relative to the aorta. Another possibility is that the patient may have breathed more deeply. We have noted movement of the needle tip associated with respiratory excursions during somatic paravertebral blocks under fluoroscopy, even when the needle shaft remained protruding the same distance from the skin. If this occurred in our case, the tip of the needle may have moved from its preaortic placement to within the aortic lumen or wall.

An argument might be raised that the use of CT imaging would have prevented this complication. This seems unlikely: radiographic contrast material could not be used because of the patient's allergy; the event occurred after successful completion of local anesthetic injection and near the end of the alcohol injection, when CT would not be in use; and fluoroscopy showed the needle to be in the appropriate location, which was verified by the location of the aortic tear found at autopsy. In addition, reliance on CT imaging was not believed to be necessary in Ischia's original description, [3]and two articles state x-ray guidance did not alter complications or outcomes. [12,13].

It is conceivable that a 22-G needle may not have caused this series of events. Celiac plexus blocks are described most often using no smaller than a 20-G needle. [3,4,6,7,9,12]The main reason is that the much greater resistance to injection through a 22-G needle makes it difficult to distinguish injecting in dense or restricting tissue compared to the loose periaortic tissue. [14]It would be more difficult to use a 22-G needle with the loss-of-resistance technique for identifying passage through the aortic wall. [4]In addition, a 20-G needle would minimize deflection from the intended path and bending when passing through dense or muscular tissue. A 22-G needle might also be more difficult to pass through a muscular aorta.

Translumbar aortography, which punctures the lumbar aorta, has a mortality of 0.05%, with 27% of fatalities resulting from aortic dissection and aneurysm rupture. [15]This would be an incidence of 1.35 per 10,000 procedures. Based on this low incidence and that reported by Davies, [5]only large series or metaanalysis of many reports would be able to determine a more precise incidence of aortic injury during transaortic celiac plexus block. In addition, dissection of the superior mesenteric artery, requiring emergency bypass surgery, has been reported with aortic puncture during translumbar aortography. [16].

Hepatic artery dissection is a rare occurrence. Larson et al. [17]tabulated and summarized their own and ten other reports from the literature. The available clinical data from these cases showed hypertension was not present in 3 of 10 cases, intramedial hepatic artery disease was absent in 3 of 10, intraabdominal manipulation did not occur in 7 of 11, and atherosclerotic arterial disease was absent in 7 of 11. Although our case did not have hypertension, intramedial arterial disease, tumor involvement, or intraabdominal manipulation, the cases tabulated by Larson et al. [17]indicate that these need not be present for hepatic arterial dissection to occur.

Finally, does the transaortic approach offer an advantage over other techniques? The high success rate of celiac block in relieving abdominal pain does permit analysis of a smaller number of patients regarding the efficacy of this approach. Ischia et al. [8]showed in a prospective randomized study that there was no significant difference in outcome between three different approaches to neurolytic celiac plexus block. The transaortic approach therefore must be reevaluated in light of our case and that of Sett and Taylor, [11]its failure to prevent paraplegia, [10]and the known mortality and complications experienced from translumbar aortography. Although complications will potentially occur with neurolytic celiac blocks, [5,18]the technique chosen should have the least chance for mortality.

In summary, we present a case of transaortic neurolytic celiac plexus block that led to the patient's demise from aortic dissection associated with needle laceration of the anterior aorta. This is the second case reported of aortic injury associated with celiac neurolytic blocks and the first report of aortic injury using the transaortic approach.

The authors thank Dr. Michael Marin, Division of Vascular Surgery, Montefiore Medical Center, for his input regarding this manuscript.