Tunneled Percutaneous Transesophageal Gastrostomy for Treatment of Malignant Bowel Obstruction

Malignant bowel obstruction (MBO) is a complication of advanced intra-abdominal malignancy estimated to affect 3–15% of patients with cancer^[1] and up to 50% of patients with advanced ovarian cancer.^[2] For many of these patients, MBO has serious impacts on quality of life with difficulty managing symptoms of nausea, vomiting, and abdominal pain. Many of these patients are not surgical candidates and the estimated life expectancy for inoperable MBO is 4–5 weeks with a 6-month life expectancy of 8% for these patients with inoperable cancer. For these patients, gastric decompression with aspiration via a nasogastric tube or percutaneous gastrostomy can help relieve symptoms.^[1]

However, certain patients may not be appropriate candidates for percutaneous transabdominal gastrostomy tube placement. Percutaneous transesophageal gastro-tubing or gastrostomy (PTEG) is a procedure alternative to a percutaneous transabdominal gastrostomy tube initially described in 1994 in Japan.^[3,4] Since then, several studies have looked at the safety and efficacy of PTEG for feeding and gastrointestinal decompression.^[5–13] PTEG is useful for patients with contraindications to percutaneous gastrostomy such as ascites, intra-abdominal metastases anterior to the stomach, gastric cancer particularly if it involves the anterior gastric wall, or intervening bowel between the stomach and anterior abdominal wall. In a 2013 prospective, multicenter study comparing nasogastric tube versus PTEG decompression, a PTEG was more comfortable than a nasogastric tube with improved relief of symptoms. This tube is often used with a suction pump for decompression.^[14] The use of a portable suction pump or suction drainage bag allows for patients to decompress at home or a facility that may not accept nasogastric tubes. This is especially important for patients with cancer who may be discharged to hospice. For many patients with malignant bowel obstructions, PTEG decompression may also allow them to have some level of oral intake for comfort.^[11]

Within surgical and interventional radiology fields, the process of tunneling central venous catheters has been used since 1967, when central catheters were initially tunneled to provide long-term vascular access for parenteral nutrition.^[15] Although today many dedicated tunneled lines have a Dacron cuff, it is still possible to tunnel noncuffed catheters as described by Sasadeusz et al^[16] as a replacement of PICC lines for patients undergoing hemodialysis. Using this tunneling technique, PTEG catheters were tunneled along the left chest wall for improved patient comfort and esthetic appearance.

A retrospective review of all institutional PTEG tubes placed between Dec 22, 2017, and Oct 3, 2023, was performed. This study was deemed exempt from review by the institutional review board and informed consent was waived, as patient data were deidentified prior to dissemination. Chart review was performed for all PTEG procedures to assess patient demographics including age at the time of placement, sex, and patient-reported race or ethnicity; procedural details such as preprocedural antibiotics, tunneled versus nontunneled tube placement, type of sedation, length of intraprocedural time, fluoroscopic time, and fluoroscopic dose length product; postprocedural outcomes including complications, symptom relief, maximum diet, and time to discharge; and overall outcomes including patient mortality and time to death.

PTEG tube placement has been previously described in detail, often using a PTEG kit (Sumitomo Bakelite, Tokyo, Japan),^[4,5] although other articles have described PTEG tube placement using readily available materials in many interventional radiology suites.^[7,10,13] All patients were required to demonstrate improvement of symptoms (most often abdominal pain, nausea, and/or vomiting) with nasogastric decompression prior to PTEG tube placement. At our institution, a 20 × 40-mm Atlas balloon (BD, Tempe, AZ) is inserted over an 0.035” stiff Amplatz guidewire (Boston Scientific, Marlborough, MA) to the upper esophagus and the balloon is inflated with a mixture of normal saline and contrast. The balloon is then pulled beneath the cricopharyngeus (Fig. 1A). An ultrasound probe is placed over the left neck to identify the balloon and surrounding structures including the trachea, left thyroid lobe, and left carotid artery. A safe tract avoiding the left carotid artery is identified and using ultrasound guidance, an 18-gauge needle is used to puncture the balloon (Fig. 1B). A second stiff 0.035” Amplatz guidewire is then inserted through the needle and advanced with the deflating Atlas balloon catheter into the stomach under fluoroscopic guidance. The balloon catheter from the nose is then released from the guidewire and removed.

Attention is then turned to the left chest, and local anesthetic is used to create a tract from the needle insertion site to an incision site a few inches away. A 12–14 Fr × 60-cm pigtail drainage catheter is then prepared with a metal stiffener and tunneled approximately 3–5 cm along a subcutaneous tract to the needle insertion site in the left neck. The metal stiffener is removed, and the catheter is advanced over the guidewire in the left neck, through the esophagus, and into the stomach. The guidewire is removed, and the catheter is sutured into place on the left chest wall. The final appearance of a traditional nontunneled PTEG catheter versus the newly described tunneled PTEG catheter is shown in Figure 2.

Descriptive statistics of patient characteristics and outcomes were collected and calculated using Excel (version 2405, Microsoft, Redmond, WA). Adverse event classification was determined according to the Society of Interventional Radiology Adverse Event Classification.^[17]

A total of 27 patients had PTEG tube placement attempts with 26 PTEG tubes placed during the study period. One PTEG tube placement failed owing to retrotracheal positioning of the esophagus, which limited ultrasound visualization. This patient was not included in the analysis of postprocedural outcomes. PTEG tube placement was attempted in 18 female and 9 male patients. Patient demographics are further delineated in Table 1.

Twenty-six of 27 PTEGs were indicated for venting of malignant bowel obstruction. One PTEG tube was placed for feeding in a patient with prior near-total gastrectomy with failed jejunostomy feeds, necessitating PTEG tube placement for tube feeds. This patient was included in the analysis to assess procedural factors and complications but was not included in the assessment of postprocedural diet or mortality outcomes given the absence of MBO. Malignancies within the patient population are listed in Table 1. Of the patients with MBO, 24 patients had ascites and 24 patients had peritoneal metastases, limiting percutaneous transabdominal gastrostomy tube placement. One patient with MBO had no peritoneal metastases or ascites but instead had dilated loops of small bowel overlying the stomach, which did not resolve with nasogastric tube decompression.

Ten tunneled and 16 nontunneled PTEG tubes were placed. Twenty-two tubes were 14-Fr and four were 12-Fr. Twenty-four PTEG tubes were placed within the stomach, one was placed postpylorically within the duodenum with additional side holes within the stomach to allow venting of the proximal small bowel and stomach, and one was placed in the jejunum in a patient with prior partial gastrectomy requiring tube feeds. Twelve patients did not receive preprocedure antibiotics. Preprocedural antibiotics were at the discretion of the primary operator, and the regimens used are shown in Supplemental Table S1, available online. Sixteen procedures were performed with moderate sedation and 11 procedures were performed with general anesthesia. The median procedure time was 48.0 minutes (± 25.0 minutes) with a median fluoroscopy time of 5.8 minutes (± 3.8 minutes) and a median fluoroscopic dose length product of 8.0 Gy cm² (± 84.9 Gy cm²).

Eight patients reported partial relief of symptoms after PTEG tube placement, and 16 patients reported complete relief of symptoms. One patient who had a PTEG tube placed as an outpatient did not have follow-up notes to document relief of symptoms or diet achieved. Seventeen patients were able to achieve a liquid, low-fiber, or soft regular diet after PTEG tube placement. The maximum diet achieved for all patients is shown in Figure 3.

Complications after PTEG tube placement are shown in Table 2, with an overall complication rate of 37.0% (10 of 27 patients with attempted placements). Most complications were minor or moderate, with two severe adverse events requiring escalation of care to the intensive care unit (ICU) after PTEG tube placement.¹⁷  One of these complications was due to difficulties extubating after PTEG tube placement, thought to be related to hypoxemia from preexisting pleural effusions. The other ICU admission was for hypotension requiring pressors immediately post procedure. Ten of the 13 postprocedural adverse events (not including failed initial placement) occurred in tunneled PTEG tubes, whereas 3 of the 13 postprocedural adverse events occurred in nontunneled tubes. Of note, all the postprocedural site infections occurred in tunneled PTEG tubes. One of these tunneled PTEG infections required long-term treatment with antibiotics without resolution of the site infection, ultimately leading to un-tunneling of the PTEG tube.

Twenty-four patients had died at the time of this writing after PTEG tube placements. The median time between PTEG tube placement and death was 24 days (± 26.6 days). One patient no longer required a PTEG tube for symptom management and the PTEG tube was successfully removed without adverse event 187 days after placement. Two patients remain alive with indwelling PTEG tubes at the time of this writing, with respective post procedure times of 120 and 154 days after PTEG tube placement. Of note, one of these patients does not have MBO and a PTEG tube was placed for tube feeds rather than symptom relief.

PTEG tube placement is an alternative for feeding and gastric decompression in patients with MBO and contraindications to a traditional percutaneous gastrostomy tube. Previously described techniques suture the PTEG catheter at the neck. Securement of the catheter in this location is sometimes bothersome to patients.^[4] Tunneling of small-bore catheters has been described as a way to move the port lumens further from a patient’s face, a technique most commonly used for long-term intravenous catheters.^[16] By combining these two common procedural methods, a PTEG catheter tunneled along the left chest wall can improve patient comfort in a similar manner.

The total complication rate in this series is higher than in prior studies, which have previously described complication rates of 22.5–23.5% with PTEG tube placement,^[4,11] although only two major complications were observed in our series, at least one of which was related to underlying patient factors limiting postprocedural extubation. Tunneled and nontunneled PTEG catheters expectedly had similar functions, with patients in both groups able to achieve gastric decompression and symptom relief. However, more complications were noted in tunneled catheters than nontunneled catheters. In particular, all site infections were noted in tunneled PTEG rather than nontunneled PTEG catheters. One of these infections did not resolve despite long-term antibiotics, and ultimately this tunneled PTEG tube was converted to a nontunneled PTEG tube to help treat the tract infection. This increased risk of infection may be related to a longer length of superficial soft tissues exposed to reflux of esophageal contents, leading to contamination and infection of the subcutaneous track, though further research would be useful to assess this trend. Purposely built cuffed PTEG tubes may limit complications related to tube displacement and dislodgement. Cuffed catheters decrease the risk of infection for tunneled central venous catheters and could have a similar effect for a tunneled PTEG tube; however, this benefit may not occur for PTEG tubes, as the esophagus is not a sterile lumen compared to a jugular or subclavian vein.

Most patients who were able to tolerate a diet for comfort were only able to tolerate a liquid diet. This is likely related to the degree of bowel obstruction and the ability to suction only consumed liquids out through the PTEG tube. A few patients were able to tolerate a diet of puree or soft foods; this is favored to be related to a lower-grade bowel obstruction with some forward bowel movement of the consumed foods. For this patient population with MBO and often limited life expectancy, PTEG tubes can be helpful to improve quality of life and allow patients some oral intake for comfort.

This series is limited by its small sample size and retrospective nature, which limits the comparison of the adverse events and outcomes of each group. Owing to the retrospective nature, only qualitative measures of symptom relief could be assessed, and some patients did not have complete postprocedural follow-up available for review.

Consistent with prior studies, PTEG tube placement is an important alternative for symptom relief from MBO in patients who are not candidates for percutaneous gastrostomy placement. PTEG catheters can be tunneled beneath the skin similar to tunneled central lines to improve patient cosmesis. Tunneling had expected similar function and may improve patient comfort. However, tunneled PTEG tube placement was complicated by more adverse events, particularly increased soft tissue site infections, compared to nontunneled PTEG catheters. As PTEG tube placement becomes more widely adopted around the world, future studies could continue to compare the potential differences between nontunneled and tunneled catheters, potentially trialing a cuffed catheter in the future.