Epidural pain management in advanced gastric cancer with inoperable malignant bowel obstruction

Background

Bowel obstruction secondary to malignancy, referred to as malignant bowel obstruction (MBO), occurs in 5%–13% of all patients with cancer.1 MBOs are most commonly associated not only with cancers of the colon and ovaries but also occurs in 6%–19% of patients with stomach malignancy.1 Mechanisms of obstruction vary and include mechanical obstruction from a physical blocking of transit and functional obstruction from an inhibition of the physiological peristaltic motions of the gastrointestinal tract. Specifically in MBO, obstruction presentations can be due to mechanical obstruction from mass(es), mechanical obstruction from adhesions due to previous surgical resection attempts, functional obstruction secondary to metastatic infiltration of the enteric nervous system, functional obstruction due to the heightened systemic inflammatory mediators with cancer and functional obstructions due to the medical treatments necessary to manage the cancer and its symptoms.1 2 Unfortunately, mean survival in patients with inoperable MBO is 4–5 weeks.1

MBO management is difficult; many mechanical obstructions are non-operable and functional treatment options are limited. Furthermore, patients with MBO suffer from significant pain, including visceral pain from the malignancy, neuropathic pain from malignant neural invasion and visceral pain from enteric organ distention proximal to the obstruction. Analgesia in these patients, usually in their last months of life, is essential but challenging. Typically, pain is addressed in MBOs with opioids. However, in scenarios where there are multiple contributors to pain, a prolonged inoperable obstruction not responding to medical management, and pain poorly controlled with hundreds of oral morphine equivalents, non-opioid epidural pain management can be an underutilised addition. Though not commonly a part of non-operative pain management, epidural analgesia is commonly regarded as a part of interventional pain approaches, the unofficial ‘fourth step’ in the WHO analgesic ladder for pain management.3 In MBO, an epidural offers a potential alternative to provide analgesia while decreasing opioid use, which may contribute to functional obstruction by slowing gut motility.3

Case presentation

A man in his 20s with stage IV metastatic gastric adenocarcinoma presented with 2 weeks of diffuse abdominal pain, low back pain, constipation and generalised weakness. On presentation, he described 8/10 diffuse abdominal pain that was both aching and sharp. He also endorsed constipation, nausea, emesis and aching back pain but denied obstipation. He was 1 day post-discharge from an outside hospital where he had been admitted for 10 days for conservative management of partial small bowel obstruction (SBO). Prior to his discharge, he underwent gastrograffin test, which showed contrast advancing to the sigmoid colon but suggesting ‘at least’ a partial SBO. Shortly after discharge, he attempted to eat but experienced severe 10/10 pain and nausea, prompting his presentation to our emergency department. In the emergency department, the patient’s physical examination was notable for being cachectic with tense abdominal distention, diffuse tenderness to light palpation, no guarding or rebound tenderness, and minimal bowel sounds.

Of note, the patient’s gastric adenocarcinoma was diagnosed 5 months prior at an outside hospital when he presented with a bowel obstruction. The obstruction was found to be secondary to a tumour at the splenic flexure, requiring exploratory laparotomy and partial colectomy. Biopsy confirmed gastric adenocarcinoma with metastasis to a perineal nodule and to the splenic flexure, which was resected due to acute bowel obstruction. Following his colectomy, the patient continued to have symptoms of abdominal pain, nausea, vomiting and constipation. In the 2 months prior to his current presentation, he experienced five admissions for partial SBO, all managed non-operatively. Of note, one of the CT scans from a recent admission documented a ‘soft tissue prominence with matting of the small bowel in the right lower quadrant, which appears to be a point of obstruction … These may be related to adhesions vs peritoneal metastasis’. Other recent CTs noted no focal transition points. During periods between obstructions, the patient reported little to no pain.

Investigations

Initial laboratory tests were significant for an elevated C reactive protein (16.04 mg/L), normocytic anaemia (haemoglobin 10.9 g/dL, mean corpuscular volume 85.2 FL) and elevated alkaline phosphatase (249 IU/L). He had normal lactic acid (1.3 mmol/L), white cell count (5.4×10⁹/L), lipase (56 U/L), basic metabolic panel and urinalysis.

CT of the abdomen and pelvis with intravenous contrast showed dilated gastric wall consistent with the patient’s gastric adenocarcinoma, moderate volume ascites and dilated loops of small and large bowel similar to the patient’s prior CT on recent admission, without a focal transition point, concerning for partial SBO or adynamic ileus (figure 1).

Differential diagnosis

The patient presented with signs and symptoms consistent with a bowel obstruction, including abdominal pain and distention with lack of bowel movements. This was similar to his recent admissions for partial SBOs. The possible aetiology of the obstruction included a combination of mechanical obstruction secondary to a metastatic mass, mechanical obstruction secondary to adhesions from previous partial colectomy, functional obstruction secondary to direct infiltration of the enteric nervous system by a mass, functional obstruction secondary to inflammation from the malignancy or functional obstruction secondary to opioid side effects.

Treatment

Initially, the patient refused nasogastric tube (NGT) placement, so management included intravenous fluids, non per os (NPO), assessment by the surgical team, initiation of pain management (acetaminophen, duloxetine, hydromorphone) and initiation of a bowel regimen (polyethylene glycol, senna). The surgical team recommended medical management as the CT did not show complete bowel obstruction or transition point. On hospital day 3, following worsening distention and nausea along with continued abdominal pain, the patient agreed to NGT placement with low intermittent suction for bowel decompression. Immediate NGT output was 500 mL. His bowel regimen was expanded (metoclopramide, dronabinol, lactulose, enemas), and analgesia was escalated to patient-controlled analgesia (PCA) pump of hydromorphone of 0–0.2 mg/hour basal rate with as needed 0.3 mg every 10 min and a few days later, a fentanyl patch was added (initially 25 μg/hour, titrated up to 125 μg/hour).

By hospital day 6, there was still no clinical improvement and high dose pain medications were needed, equating to over 600 OMEs per day. The patient was more somnolent, though easily arousable, and NGT output in the 3 days since its placement averaged over 900 mL per day. Abdominal X-rays during this time period showed no change in bowel distention pattern. The team hypothesised that the high opioid requirement was contributing an opioid-induced ileus atop of the underlying malignant obstruction, and, therefore, explored other pain control options. Gastroenterology and interventional radiology were consulted and concluded he would not benefit from a venting gastric tube. Sublingual buprenorphine was given in attempt to reduce constipation while providing adequate pain control. Though this regimen produced a small bowel movement, the patient’s pain immediately increased. Anaesthesia and interventional radiology were consulted for a possible coeliac plexus block. It was determined that the possibility of mechanical obstruction due to adhesions or masses was a contraindication to a plexus block, but the anaesthesia team offered a thoracic epidural instead. The patient declined. Given his inability to tolerate oral nutrition due to nausea and vomiting, parenteral nutrition was initiated on hospital day 8. As another attempt to decrease the opioid contribution to his obstruction, subcutaneous methylnaltrexone was given to antagonise the peripheral opioid effects, which resulted in a small bowel movement without change in distention or pain, leading the team to believe an opioid-induced ileus was now a contributing factor to his obstruction. On hospital day 10, an abdominal X-ray revealed new linear lucencies along the small bowel wall, for which the radiologist could not rule out pneumatosis. This finding triggered a repeat abdominal CT, which showed small bowel dilation up to 6 cm and a transition point in the right lower quadrant, but no pneumatosis and ‘no evidence of metastatic disease in the abdomen or pelvis’, leading the team to believe adhesive disease was contributing to the obstruction (figure 2).

On hospital day 11, the patient was agreed to an epidural, which the anaesthesia team placed at the T7–8 level. Ropivacaine 0.2% infusion was begun at a basal rate of 6 mL/hour with a patient bolus of 4 mL available every hour. With the epidural infusion and fentanyl patch, the patient did not require any PCA doses with hydromorphone. Decadron and octreotide were added and within the next 24 hours, the patient had three large bowel movements with improvement of his pain.

Outcome and follow-up

Unfortunately, on hospital day 14 (the fifth day with the epidural), the patient developed fevers, chills and myalgias. Empiric treatment with intravenous vancomycin, cefepime and metronidazole was started for sepsis with a suspected abdominal source. His blood cultures grew bacillus species in 4 of 4 bottles. The epidural was removed given his bacteremia, and a 10-day course of vancomycin was continued.

An aggressive bowel regimen was continued and his pain was ultimately controlled with methadone and oxycodone. He tolerated a liquid diet, the NGT was clamped then removed, and parenteral nutrition was discontinued. On hospital day 24, the patient was discharged with palliative care referral based on patient preference.

Medical records indicate that the patient has continued to receive palliative treatment until the time of this report submission, more than 6 months after the described encounter. Records also indicate three additional admissions for recurrent partial SBO, all managed non-operatively.