Placement of a feeding tube through a gastrostomy can be performed endoscopically or radiologically. While percutaneous endoscopic gastrostomy (PEG) tube placement is most frequently performed using a "pull" technique, this method may not feasible in patients with malignant, or tight benign, esophageal stenosis. Further, the "pull" technique may drag tumor cells with the feeding tube and lead to implantation metastasis at the gastrostomy site. A clinical practice update by the American Gastroenterological Association has recommended that the pull-through PEG placement method should be avoided in all patients with oropharyngeal or esophageal cancer. It also recommends that the introducer/Push PEG method should be favored instead of the pull PEG. In such situations, an introducer-style, "Direct" gastrostomy tube can be placed endoscopically or radiologically. However, the published data comparing outcomes and safety of endoscopic "Direct" PEG (D-PEG) and interventional radiological PEG (IR-PEG) are very sparse.

The D-PEG is performed under endoscopic visualization of the gastric wall which facilitates greater control and allows safe selection of gastrostomy site. Further, the presence of an endoscope enables transillumination to confirm the absence of intervening abdominal viscera between the abdominal wall and the anterior wall of the stomach. These advantages are lacking with the IR-PEG. We hypothesize that D-PEG is safer than IR-PEG.

In this single center, non-randomized study, patients unable to undergo a conventional per-oral "Pull" PEG and needing a D-PEG will be prospectively enrolled. For the comparison arm, historical IR-PEG procedures at our center will be assessed. The technical success and rates of adverse events will be compared between the two arms. Approval from the Institutional review board has been obtained. Based on our experience, we estimate a sample size of 40 participants in each arm and anticipate completion of this pilot study by June 2021.

Introduction Percutaneous endoscopic gastrostomy (PEG) tube placement is a frequently performed procedure to provide enteral nutrition in patients with dysphagia. Two methods are used for placing a PEG tube via the mouth: The "Pull" method and the "Push" method(1). In the "Pull" technique, the feeding tube is pulled from the mouth, through the esophagus and stomach to the outside. In the introducer-style "Direct" technique, the feeding tube is pushed from the outside to inside the stomach. Both these techniques are considered safe, effective, and have their own advantages and disadvantages(2).

Limitations of Pull PEG technique The "Pull" PEG technique is challenging in patients with head and neck malignancies, obstructing esophageal cancers, and laryngeal malignancies. The severe stenosis often precludes the passage of the adult endoscope as well as the large internal bumper of the feeding tube which must be pulled through the stenotic upper gastrointestinal tract. Further, microorganisms and malignant cells from the oropharynx/ esophagus can be pulled by the feeding tube leading to peristomal infections as well as metastatic deposits at the site of the gastrostomy(3,4). A prospective evaluation of "Pull" PEGs placed in patients with oropharyngeal and esophageal tumors revealed metastatic deposits of malignant cells in 9.5% of patients, 16 weeks after the PEG(5). Retrospective assessment of PEG placement for oropharyngeal cancer demonstrated abdominal wall metastasis of 0.64%(6).

Direct PEG technique The endoscopic Direct PEG (D-PEG) uses the Russel introducer technique which obviates the passage of the gastrostomy tube through the oropharynx. Further, it can also be performed with an ultrathin endoscope which is easy to navigate through stenotic or malignant upper gastrointestinal tract obstructions. The D-PEG requires performance of gastropexy (attachment of gastric wall to the abdominal wall using sutures) with anchor-like T-fasteners for easy and safe placement of the PEG tube(2).

The D-PEG is performed under endoscopic visualization of the gastric wall which facilitates greater control and allows safe selection of gastrostomy site. It also obviates the need for ionizing radiation. Further, the endoscope enables transillumination to confirm the absence of intervening abdominal viscera between the abdominal wall and stomach. These advantages are lacking in IR-guided PEG which does not offer endoscopic visualization of the gastric lumen.

Currently, patients who fail a per-oral Pull PEG tube placement are typically referred for IR-guided PEG placement(3). Thus, patients who fail a "pull" PEG procedure, are referred to IR for a second invasive procedure. However, with the availability of the D-PEG, patients failing a "Pull" PEG can undergo a PEG placement during the index procedure.

Recommendations from GI societies A recently published clinical practice update has recommended that the pull-through PEG placement method should be avoided in all patients with oropharyngeal or esophageal cancer, to reduce the risk of implantation metastasis. Further, the authors recommend that the introducer/Push PEG method should be favored instead of the pull PEG(7,8).

Hypothesis We hypothesize that the D-PEG is safer than the IR-PEG in patients unable to undergo a per-oral Pull PEG.

Aims Primary Aim: Comparison of adverse events of D-PEG vs IR-PEG in patients unable to undergo a Pull PEG Secondary aims: Comparison of technical success and overall procedure duration between D-PEG and IR-PEG in patients unable to undergo a Pull PEG The data from this pilot study will be used to design a multi-center prospective randomized trial comparing outcomes and safety of D-PEG and IR-PEG. The subsequent trial is intended to be performed at different centers where these procedures are routinely performed, such as Virginia Commonwealth University, University of Utah, and Baylor College, TX.

Pilot data The investigators have received approval from the institutional review board. We have performed D-PEG procedure in 13 patients so far with technical success in 12 patients (92%). There have been no major adverse events including perforation, significant procedure-related bleeding, or procedure related death. One patient accidentally dislodged his tube which was replaced without endoscopic intervention (minor adverse event: 1).

Research design This will be a prospective study of participants undergoing D-PEG placement compared with patients who previously underwent IR-PEG placement.

Endoscopic/prospective arm The patients who fail the endoscopic pull PEG placement, or are not candidates for a Pull PEG, will undergo an D-PEG procedure at the time of the index endoscopy. During the pre-procedure consent process, participants will be offered IR-guided PEG placement as an alternative. This is the standard of care for patients who present to the endoscopy unit for PEG tube placement. All standard protocols including pre-procedure assessment, intra-procedure antibiotic use and post procedure protocols will be observed. In addition to the standard care which all patients receive, the technical success and adverse event rates will be recorded. Participants will receive a phone call within 1 week, and within 30 days of the procedure to assess for any post-procedure complications. These data will be recorded in the medical chart.

The study will be performed at Kansas City VA Medical Center, Kansas City, Missouri. The procedures will be performed by trained, board-certified gastroenterologists with extensive experience in placing PEG tubes.

Statistical Analysis/power analysis:

We will compare the technical success of D-PEG versus IR-PEG placement. Technical success is defined as the successful placement of the D PEG.

Primary outcome (adverse event rate):

We assume 15% complications (bleeding, infection, tube dislodgement, perforation) with IR-PEGs and 5% complications in D-PEGs. With 80% power to test a noninferiority margin of 5%, we intend to enroll 50 patients in the prospective arm done by endoscopists and 50 patients in the retrospective arm done by IR.

Secondary outcome (technical success) We assume a 95% technical success rate on IR guided push PEG placement. We also assume a 90% technical success rate on endoscopic push PEG placement. Using a non-inferior to margin of 20%, and an 80% power to test the hypothesis, we intend to enroll 50 patients in the prospective arm and 50 patients in the retrospective arm.

We will use Student's t-test compare means and chi-square test to compare the two arms. We will also use multi variate analysis and ANOVA for the analysis of all variables.

Recruitment of subjects

Patients with dysphagia requiring PEG placement will be recruited from the following sites:

Referred to the GI clinic for placement of a gastrostomy tube Referred by the Kansas City VA tumor board for placement of a gastrostomy tube (e.g. head and neck cancers and esophageal malignancies) Outpatient or Inpatient GI consults for gastrostomy tube placement Patient's meeting the inclusion and exclusion criteria listed above will then be assessed by the principal investigator and offered enrollment into the study.

Inclusion Criteria:

• Placement of PEG tube for dysphagia
• Inability of participant to undergo conventional Pull PEG (for any reason)
• Attempt at placement of a Push PEG
• Age > 18 years and able to consent

Exclusion Criteria:

• Successful placement of Pull PEG
• Ascites, pregnancy