A 62-year-old male was referred to the pain clinic by his oncologist for discussion of pain management options. The patient had presented 5 months earlier with a 15 lb weight loss and epigastric pain radiating through to the back. The evaluation established a diagnosis of adenocarcinoma of the head of the pancreas, and staging revealed metastases to several locations in the liver and lungs and multiple locations in the peritoneum. CT scan showed that the primary tumor mass was relatively well confined to the pancreas, with only a single localized area of enlargement of periaortic lymph nodes, and no involvement of major vessels. The patient was relatively functional, with a Karnofsky score of 80. There had been slight decrease in tumor size with three courses of gemcitabine/paclitaxel therapy, but the patient had decided against further chemotherapy because of intolerable side effects. Past medical history included hypertension and a 40-pack-year history of cigarettes.

On first presentation to the pain clinic, the patient complained of two distinct types of abdominal pain. A diffuse and vague pain throughout the abdomen, intensity 4 on a scale of 0–10, was slightly worse after meals and sometimes improved after bowel movements. A second, more intense epigastric and right upper quadrant pain, 7–8 on a scale of 0–10, has reliably worsened for 1–2 h after each meal and radiated seemingly directly through the body to the central mid back. The patient had tried hydrocodone, oxycodone, and hydromorphone with partial relief and was now taking extended-release morphine sulfate, 60 mg orally every 8 h, duloxetine, and gabapentin. This regimen provided some relief of pain but had caused unacceptable somnolence, intermittent confusion, and constipation.

On physical examination, the patient appeared intelligent and interactive but chronically ill. He was in apparent chronic distress from pain. Abdominal exam revealed normal bowel sounds, mild distention, a sense of fullness in the right upper quadrant of the abdomen without discrete palpable masses, and marked tenderness over the epigastrium. Light percussion of the central mid back was not tender, but moderate percussion was clearly tender.

A wide range of treatment options was discussed. The patient elected to proceed with a therapeutic trial of celiac plexus block with bupivacaine.

It was decided to use a two-needle, posterior, percutaneous anterocrural approach under fluoroscopic guidance [1]. With the patient prone, light IV sedation, a 22 gauge, 15 cm Chiba needle from a skin entry point just below the 12th rib, 8–10 cm lateral to the midline, was directed medially toward the L1 vertebral body. On lateral view, each needle was advanced a distance of 2–3 cm anterior to the L1 vertebral body, palpably grazing that body as the needle passed. Careful intermittent aspiration for blood was performed to avoid the aorta and vena cava. A test dose of iohexol (Omnipaque^®) 300 mg/mL, 5 mL through each needle, showed good retroperitoneal spread without vascular uptake. Once in position, an injection was made of 20–30 mL 0.25% bupivacaine through each needle, in 5 mL increments.

The patient tolerated the procedure well. Within 5 min after completion of the injection, he reported a marked improvement in pain, from 7/10 preprocedure to 1/10 post-procedure. After a period of observation, he was discharged home with his caregiver wife. Over the next several days, the patient was happy with the results of the block, although pain relief was incomplete. Intensity of the epigastric pain was down to 3/10. At approximately 72 h post-procedure, however, the pain returned to its full intensity. The patient wished to proceed with a more permanent neurolytic block, as had been previously discussed at the initial clinic visit.

In an attempt to improve efficacy of epigastric pain relief, the technique was changed to a transaortic technique [2, 3]. The approach was similar to that for the first block, modified as a single needle, left-sided approach only, intentionally directed toward the posterior wall of the aorta and then advanced through the aortic wall to the anterior side, using palpation of arterial pulsation, aspiration for blood, and intermittent injection of 1–2 mL boluses of iohexol as guides. The planned injection was for 40 mL total volume of equal parts absolute alcohol and 0.5% bupivacaine, admixed, using an intermittent aspiration technique. After approximately 30 mL was injected, blood was aspirated through the needle. The needle was advanced to reposition it anterior to the aorta. When pressure was exerted on the plunger of the syringe, the patient complained of a transient unusual abdominal and lower thoracic pain, and blood was again aspirated through the needle. The pain appeared to subside within seconds. The needle was again advanced to a position likely anterior to the aorta by AP and lateral fluoroscopic view; an iohexol test dose showed good position on AP and lateral fluoroscopy, and the remaining 10 mL of alcohol/bupivacaine was injected.

The patient was observed for 4 h in the recovery room with stable vital signs, normal urination and ambulation, and a marked decrease in his preprocedure pain. He was discharged home with his wife.

Four hours post-procedure, the wife reported by telephone that the patient was experiencing diaphoresis, severe abdominal pain, and inability to move the legs. The paramedics returned the patient to the hospital with vital signs that became increasingly unstable en route. Consistent with the patient’s documented wishes, advanced resuscitation was not performed, and the patient expired 30 min after arrival in the emergency room.

A coroner’s autopsy was performed. The abdominal aorta had an intramural dissection of the anterior wall between outer media and adventitia layers, centered at the level of L1 and extending proximally and distally, forming a pseudoaneurysm and thrombus with occlusion of a major anterior spinal artery branch at T12 and occlusion of the celiac axis and all of the distal aortic branches. There was ischemic death of a segment of the spinal cord at T12, as well as most of the organs of the abdomen. There was moderate intraluminal plaque in the aorta.

Initial evaluation: Avoidance of complications can start before the initial interview, by screening for inappropriate referral. Conversation with the referring physician can clarify life expectancy and family dynamics. Life expectancy greater than 3 months may mean repeat block will be needed [4]. A life expectancy of only a few weeks may suggest insufficient benefit to the patient to match the risk and inconvenience of undergoing the procedure.

The approach strategy to celiac ganglia may be affected by tumor. For example, a celiac plexus heavily infiltrated by tumor mass may suggest a more proximal, retrocrural splanchnic approach. Prior abdominal CT scan will often be available, and reviewing abdominal imaging as part of planning the procedure is useful. At times, however, imaging may be several months old, and the patient’s advanced state of disease may not warrant additional imaging.

The block is likely to help only the pain from the structures innervated by the celiac plexus [5]. In this case example, pain caused by distal colon, lung, pleura, and peritoneal metastases would not be expected to improve. Thoughtful diagnosis of specific causes of the several pain elements a patient may be experiencing is almost always valuable to the practitioner and patient.

It is also important to assess patient and family goals. The referring physician or independent patient/family research may have created unrealistic expectations. During the initial patient consultation, it is important to be supportive but modest about the benefits of celiac plexus block. There is a reasonable likelihood of significantly decreasing pain. The block cannot promise improved quality of life or prolongation of life expectancy [6]. Common possible complications should be discussed, including the possibility that the block will be ineffective.