1

Introduction

Nutritional support is recognized as a vital component of head and neck cancer (HNC) treatment since malnutrition is a risk factor for post-operative complications and poor outcomes . HNC patients commonly present with malnutrition and weight loss as a result of difficult mastication or dysphagia . Other HNC patients will develop dysphagia as a result of treatment-related mucositis, edema or fibrosis .

Percutaneous endoscopic gastrostomy (PEG) is a minimally invasive method that can be used to gain durable enteric feeding access in HNC patients . PEG use in HNC has been shown to be safe and effective for delivering enteral nutrition and decreases the need for parenteral nutrition . PEG also allows patients to avoid the discomfort and social stigma associated with nasoenteric feeding tubes.

PEG tubes in HNC are removed after patients have recovered swallowing function after successful completion of initial treatment. Since some HNC patients will experience disease recurrence requiring additional treatment and nutritional support, these patients may be considered for repeat PEG placement . A repeat PEG may also be necessary to treat long-term complications of treatment such as dysphagia or aspiration, or a repeat PEG may be necessary as part of a palliative care treatment plan. There is a paucity of data regarding repeat tubes in HNC patients . We report our results of PEG in HNC patients with focus on repeat PEG by comparing results of repeat PEGs to first-time PEGs.

2

Materials and methods

After obtaining IRB approval for the study, HNC patients having undergone PEG placement at the University of Missouri between August 2009 and July 2011 were identified for inclusion. Patients who underwent open tube gastrostomy were excluded. Data collected by chart review included cancer type, initial or repeat PEG, preoperative body mass index (BMI), number of previous abdominal operations, concurrent surgical or endoscopic procedures, and PEG-related complications. Repeat PEG was defined as placement of a PEG tube in a patient who previously had a gastrostomy tube. PEG complications were defined as (1) organ space infection, (2) peritonitis, (3) intraperitoneal leak, (4) intra-abdominal organ injury, (5) gastrocolocutaneous fistula, and (6) tube dislodgement. Data regarding superficial surgical site infection were deemed too unreliable for collection.

A 20-French pull PEG kit (Boston Scientific, Natick, MA) was used in all cases. Patients underwent induction of general anesthesia followed by peroral introduction of an Olympus GIF Q180 gastroscope (Olympus, Center Valley, PA). After inspection of the esophagus, stomach and duodenum, the site for PEG placement was located by finger impulse and abdominal wall transillumination. Placement of the PEG was by pull method as previously described.

Statistical analysis was conducted with SPSS v.18® (IBM, Inc.). Student’s independent t -test was used to compare continuous data. Chi-square test and Fischer’s exact test were used as appropriate for analyzing categorical data. Odds ratios with confidence intervals are reported for comparison of complications. Regression analysis was planned, but could not be performed because of the low number of events. Statistical significance is considered at p -value < 0.05.

2

Materials and methods

After obtaining IRB approval for the study, HNC patients having undergone PEG placement at the University of Missouri between August 2009 and July 2011 were identified for inclusion. Patients who underwent open tube gastrostomy were excluded. Data collected by chart review included cancer type, initial or repeat PEG, preoperative body mass index (BMI), number of previous abdominal operations, concurrent surgical or endoscopic procedures, and PEG-related complications. Repeat PEG was defined as placement of a PEG tube in a patient who previously had a gastrostomy tube. PEG complications were defined as (1) organ space infection, (2) peritonitis, (3) intraperitoneal leak, (4) intra-abdominal organ injury, (5) gastrocolocutaneous fistula, and (6) tube dislodgement. Data regarding superficial surgical site infection were deemed too unreliable for collection.

A 20-French pull PEG kit (Boston Scientific, Natick, MA) was used in all cases. Patients underwent induction of general anesthesia followed by peroral introduction of an Olympus GIF Q180 gastroscope (Olympus, Center Valley, PA). After inspection of the esophagus, stomach and duodenum, the site for PEG placement was located by finger impulse and abdominal wall transillumination. Placement of the PEG was by pull method as previously described.

Statistical analysis was conducted with SPSS v.18® (IBM, Inc.). Student’s independent t -test was used to compare continuous data. Chi-square test and Fischer’s exact test were used as appropriate for analyzing categorical data. Odds ratios with confidence intervals are reported for comparison of complications. Regression analysis was planned, but could not be performed because of the low number of events. Statistical significance is considered at p -value < 0.05.

3

Results

The study group included 103 HNC patients in whom 107 PEG tubes were placed. Of these, 18 (17%) were placed in patients who previously had a PEG tube (repeat PEG). Cancer recurred in four patients within the study period and these patients are included in both the initial PEG and repeat PEG groups. Repeat PEGs were most commonly placed as part of the treatment plan for a HNC recurrence (12/18, 67%). The remainder of repeat PEGs was placed for dysphagia after treatment (22%), replacement after dislodgement (5%), and primary treatment in a patient with previous PEG placed for treatment of trauma (5%).

The average age of patients in this study was 61 years (range 21–91). There was no age difference between the initial and repeat PEG groups (60.9 vs 61.8 years, p = 0.407). Mean body mass index (BMI) for all PEG patients was 25.4 (range 14.3–44.7). BMI was similar between the initial and repeat PEG groups (25.9 vs 23.0, p = 0.058). Prior abdominal surgery was more common in the repeat PEG group (OR 25.9, 95% CI 5.5–121.9, p < 0.001).

Concurrent procedures were performed in 55 (53%) patients at the time of PEG or repeat PEG placement. Combined procedures included 24 (23%) curative intent oncologic resections, 20 (19%) port-a-cath placements, 6 (6%) endoscopic biopsies, 5 (5%) tooth extraction procedures, and 2 (2%) tracheostomies. There was no difference in combined procedure between initial and repeat PEG groups (OR = 0.72, CI: 0.26–2.0, p = 0.517).

Complications occurred in four patients overall (three dislodgement, one gastrocolocutaneous fistula). There were no deep or organ space surgical site infections, episodes of peritonitis or intraperitoneal leak. Two complications occurred in the repeat PEG group (dislodgement, fistula) and two occurred in the initial PEG group (dislodgement). This difference was not significant (11% vs. 2%, p = 0.131, OR 5.4, CI: 0.71–41.5). BMI in the noncomplication group was 25.3 (range 14.3–44.7) versus 27.3 in the complication group (range 18.5–37.5) ( p = 0.491). Complication rate was similar between obese (BMI > 30) and non-obese (BMI < 30) patients ( p = 0.175).

The fistula required a minilaparotomy for repair and replacement of a surgical gastrostomy tube at a new site. All tube dislodgements occurred after a mature tract had formed and were treated with replacement of a gastrostomy tube through the PEG tract. There was no peritonitis or intraperitoneal leak associated with PEG dislodgement. No mortality occurred in either group.